Localization and Positioning Systems for Emergency Responders: A Survey 应急响应人员定位系统：调查
André Filipe Gonçalves Ferreira, Duarte Manuel Azevedo Fernandes, André Paulo Catarino, and João L. Monteiro, Member, IEEE André Filipe Gonçalves Ferreira、Duarte Manuel Azevedo Fernandes、André Paulo Catarino 和 João L. Monteiro，IEEE 会员
Abstract 摘要
The availability of a reliable and accurate indoor positioning system (IPS) for emergency responders during onduty missions is regarded as an essential tool to improve situational awareness of both the emergency responders and the incident commander. This tool would facilitate the mission planning, coordination, and accomplishment, as well as decrease the number of on-duty deaths. Due to the absence of global positioning system signal in indoor environments, many other signals and sensors have been proposed for indoor usage. However, the challenging scenarios faced by emergency responders imply explicit restrictions and requirements on the design of an IPS, making the use of some technologies, techniques, and methods inadequate on these scenarios. This survey identifies the specific requirements of an IPS for emergency responders and provides a tutorial coverage of the localization techniques and methods, highlighting the pros and cons of their use. Then, the existing IPSs specifically developed for emergency scenarios are reviewed and compared with a focus on the design choices, requirements, and additional features. By doing so, an overview of current IPS schemes as well as their performance is given. Finally, we discuss the main issues of the existing IPSs and some future directions. 为执行任务的应急响应人员提供可靠、准确的室内定位系统（IPS）被认为是提高应急响应人员和事件指挥官对态势感知的重要工具。这一工具将促进任务的规划、协调和完成，并减少因公殉职的人数。由于全球定位系统信号在室内环境中的缺失，人们提出了许多用于室内的其他信号和传感器。然而，应急响应人员所面临的挑战性场景对 IPS 的设计提出了明确的限制和要求，使得一些技术、工艺和方法在这些场景中的应用变得不充分。本调查报告明确了应急响应人员对 IPS 的具体要求，并提供了定位技术和方法的教程，强调了使用这些技术和方法的利弊。然后，对专门为应急场景开发的现有 IPS 进行回顾和比较，重点关注设计选择、要求和附加功能。这样，我们就能对当前的 IPS 方案及其性能有一个大致的了解。最后，我们讨论了现有 IPS 的主要问题和一些未来发展方向。
Index Terms-Emergency responders, indoor positioning systems, indoor localization, localization methods, localization techniques, unknown and unstructured environments. 索引词条--应急响应人员、室内定位系统、室内定位、定位方法、定位技术、未知和非结构化环境。
I. Introduction I.导言
NOWADAYS, there is a huge familiarization with localization systems. This happens because they are ubiquitous in a wide range of applications in our daily life. In part, much of the familiarity with these systems is due to the Global Positioning System (GPS) receivers. Such devices are used in a vast range of applications: pedestrian and/or car navigation systems; fleet control by the companies; georeferencing 如今，人们对本地化系统已非常熟悉。这是因为定位系统在我们日常生活的广泛应用中无处不在。人们之所以熟悉这些系统，部分原因在于全球定位系统（GPS）接收器。这类设备的应用范围非常广泛：行人和/或汽车导航系统；公司的车队控制；地理参照
of photographs taken by a smartphone or tablet; military activities, just to name a few. Despite the good results provided by this technology in outdoor applications, in indoor applications the performance of GPS-based systems is unsatisfying. This happens mainly because of the weakness of GPS signals, namely incapability to penetrate the building walls and the multipath effect. 智能手机或平板电脑拍摄的照片；军事活动等等。尽管这项技术在室外应用中取得了良好的效果，但在室内应用中，基于 GPS 的系统的性能却不能令人满意。这主要是因为 GPS 信号的弱点，即无法穿透建筑物墙壁和多径效应。
To overcome the GPS limitations in indoor environments, since the beginning of this millennium a huge effort on the development of an Indoor Positioning System (IPS) has been made. One of the most popular methods for indoor localization is the fingerprinting technique [1]. The popularity of this method arises from the fact that it relies on existing Wi-Fi infrastructures, which makes it a cost effective solution. Another attractive feature of this method is that unlike traditional geometric approaches (such as trilateration and triangulation), the fingerprinting method achieves meter-level accuracy even in non-line-ofsight (NLOS) environments [1]. Based on this and other methods, several IPSs are already on the market to provide indoor location-based services (ILBS) [2]-[11]. However, none of these commercial IPSs is suitable to solve the problem of emergency responders’ location during their duty [12]-[16], as all solutions require beforehand measurements, calibration, configuration and deployment. These are unreliable prerequisites in emergency scenarios due to the nature of the indoor environments, which are typically defined as unstructured or unknown [12], [13], [17]-[22]. The indoor environments can be classified as structured or known, semi-structured and unstructured or unknown depending on the control that the IPS possesses over them [23], [24]: 为了克服全球定位系统在室内环境中的局限性，自本世纪初以来，人们在开发室内定位系统（IPS）方面做出了巨大的努力。指纹识别技术是最流行的室内定位方法之一[1]。这种方法之所以受欢迎，是因为它依赖于现有的 Wi-Fi 基础设施，因此是一种经济有效的解决方案。这种方法的另一个吸引人的特点是，与传统的几何方法（如三坐标法和三角测量法）不同，即使在非视距（NLOS）环境中，指纹识别法也能达到米级精度[1]。基于这种方法和其他方法，市场上已经出现了几种 IPS，用于提供室内定位服务（ILBS）[2]-[11]。然而，这些商用 IPS 都不适合解决应急响应人员在执勤期间的定位问题 [12]-[16]，因为所有解决方案都需要事先测量、校准、配置和部署。由于室内环境通常被定义为非结构化或未知环境[12]、[13]、[17]-[22]，这些都是应急场景中不可靠的先决条件。根据 IPS 对室内环境的控制能力，室内环境可分为结构化或已知、半结构化和非结构化或未知[23]、[24]：
Structured or Known Environment: this type of scenario is characterized by the total control of the IPS over the environment. Usually, these environments are static, i.e., do not change their shape or configuration over the time or the IPS is updated when a change occurs. Moreover, the IPS relies on up-to-date maps of the building or floorplans (e.g., Building Information Modelling - BIM), which are used to restrict the possible user movements, leading to a significant reduction of the uncertainty in the estimation of the position. Additionally, landmarks, ^(1){ }^{1} 结构化或已知环境：这类场景的特点是 IPS 对环境的完全控制。通常，这些环境是静态的，即其形状或配置不会随着时间的推移而改变，或者在发生变化时 IPS 会进行更新。此外，IPS 依赖于建筑物的最新地图或平面图（如建筑信息模型 - BIM），用来限制用户可能的移动，从而显著降低位置估计的不确定性。此外，地标、 ^(1){ }^{1}
which can assume a variety of forms (e.g., radio beacons, tags for cameras, reflectors for scanners or some special features of the building geometry), can be installed in strategic places of the building to help the localization process. The position of landmarks is known beforehand by the system. The combination of building maps with landmarks results in environmental maps that provide a great support for an IPS; 地标可以有多种形式（如无线电信标、摄像机标签、扫描仪反射器或建筑物几何形状的一些特殊特征），可以安装在建筑物的战略要地，以帮助定位过程。系统会预先知道地标的位置。将建筑物地图与地标结合起来，就能绘制出环境地图，为 IPS 提供极大的支持；
Semi-Structured Environment: this environment is a subtype of an unstructured or unknown environment. To reduce the complexity of the localization process and to increase the performance of the IPS, some assumptions related to the existence or geometry of building structures (e.g., corridors, stairs, doorways or height between floors) are made. Within the IPS these building structures are defined as landmarks; 半结构化环境：这种环境是非结构化或未知环境的一种子类型。为了降低定位过程的复杂性并提高 IPS 的性能，需要对建筑物结构（如走廊、楼梯、门口或楼层之间的高度）的存在或几何形状做出一些假设。在 IPS 中，这些建筑结构被定义为地标；
Unstructured or Unknown Environment: in this type of scenario, the IPS does not have control over the environment. An IPS operating in this type of environment does not have access to floorplans, maps, or other prior knowledge about the building structure or signal propagation conditions. Moreover, no assumptions on the existence or geometry of building structures (e.g., corridors, stairs, doorways or height between floors) are made. The information that the IPS possesses about the unknown building is the one acquired from sensors carried by the user and, if available, from waypoints ^(1){ }^{1} strategically deployed on-the-fly. This is the most challenging environment for the localization process, since the IPS has to deal with dynamic changes of the environment [23], [24]. 无结构或未知环境：在这种情况下，IPS 无法控制环境。在这类环境中运行的 IPS 无法获得平面图、地图或其他有关建筑结构或信号传播条件的先验知识。此外，也无法假设建筑物结构（如走廊、楼梯、门口或楼层之间的高度）的存在或几何形状。IPS 所拥有的关于未知建筑物的信息是从用户随身携带的传感器中获取的，如果有的话，还可以从实时战略部署的航点 ^(1){ }^{1} 中获取。这是定位过程中最具挑战性的环境，因为 IPS 必须应对环境的动态变化 [23]，[24]。
Particularly in urban/structural firefighting, the majority of existing IPSs do not take into account the challenging conditions where the emergency responders work in. Darkness, power outages, water, high temperatures, smoke, flames and noise can prevent an IPS from working [22], especially IPSs that are previously installed on the building and rely on the building grid [18]. Moreover, the weight and complexity of personal protective equipment hampers the interaction between a firefighter and a laptop/smartphone. So, to provide a good acceptance among emergency responders, especially the firefighters, it is necessary to guarantee that the IPS does not interfere with the firefighter activity and, at the same time, an intuitive interaction between the firefighter and the IPS [22]. 特别是在城市/建筑消防中，现有的 IPS 大多没有考虑到应急人员工作的艰苦条件。黑暗、断电、积水、高温、烟雾、火焰和噪音都可能导致 IPS 无法工作[22]，尤其是之前安装在建筑物上并依赖于建筑物电网的 IPS [18]。此外，个人防护设备的重量和复杂性也妨碍了消防员与笔记本电脑/智能手机之间的互动。因此，为了让应急响应人员，尤其是消防员接受，有必要保证 IPS 不会干扰消防员的活动，同时保证消防员与 IPS 之间的直观交互[22]。
Despite all the challenges-unknown environments and hazardous working conditions-an IPS is an essential tool for emergency responders [12], [13], [17], [20], [22], [25]-[29]. Both the U.S. National Fire Protection Association (NFPA) and the U.S. National Institute for Occupational Safety and Health (NIOSH) have identified the disorientation of firefighters inside buildings - due to the combination of low or no visibility (caused by dense smoke and power outages), fire intensity, destruction of escape routes and limited air supply-, as the major cause of injuries and deaths among the firefighters, while fighting urban fires [12], [20], [22], [30]. Furthermore, under these extreme conditions, the firefighters are subjected to high levels of mental and physical stress, which makes the simple task of getting out of a building 尽管存在各种挑战--未知的环境和危险的工作条件--但 IPS 是应急响应人员必不可少的工具 [12]、[13]、[17]、[20]、[22]、[25]-[29]。美国国家防火协会（NFPA）和美国国家职业安全与健康研究所（NIOSH）都认为，消防员在建筑物内迷失方向--由于低能见度或无能见度（由浓烟和断电造成）、火势猛烈、逃生通道被毁和空气供应有限--是消防员在扑救城市火灾时受伤和死亡的主要原因[12]、[20]、[22]、[30]。此外，在这些极端条件下，消防员承受着巨大的精神和身体压力，这使得从建筑物中逃生这一简单的任务变得尤为困难。
a challenging task [12], [20], [22]. In this context, the delivery of information on the current location of the emergency responders and their walked paths to the incident commander is a key factor for the mission success. This information provides situational awareness of the intervention scenario. With such information, the incident commander can monitor the current location of all units, trace the searched rooms, and decide where to send additional units. Additionally, escape routes can be generated, based on the recorded tracks of the several teams, to guide disoriented firefighters to the nearest exit or to create alternatives if the shortest path is no longer available. These tracks can also be used to send rescue teams if a distress call is received. 是一项具有挑战性的任务 [12]、[20]、[22]。在这种情况下，向事件指挥官提供应急响应人员的当前位置及其行走路径的信息是任务成功的关键因素。这些信息提供了干预情景的态势感知。有了这些信息，事件指挥官就可以监控所有单位的当前位置，追踪已搜索的房间，并决定向何处增派单位。此外，还可以根据几个小组记录的轨迹生成逃生路线，引导迷失方向的消防员找到最近的出口，或者在最短路径不再可用的情况下创建替代路径。如果收到求救信号，还可以利用这些轨迹派遣救援小组。
In this survey paper, the major contributions match with the organization of the paper. Therefore, the contributions/sections are organized as follows. Section II presents a comparison of existing surveys on indoor localization. A definition of the general architecture of an IPS, as well as, the main requirements of an IPS for emergency responders are presented in Section III. Section IV provides a comprehensive survey of techniques used for measuring distances, displacement, and heading, as well as, the localization methods and its pros and cons. A new taxonomy for the classification of the IPS schemes for emergency responders is proposed in Section V. Section VI provides a comprehensive survey of state-of-art IPS schemes for emergency responders, highlighting the strengths and weakness of each scheme. This section ends with a detailed comparison summary of the state-of-art on IPSs for emergency responders, highlighting their similarities and differences, based on design choices, requirements, and additional features of each IPS (Section VI-E). The issues related to the current IPSs for emergency responders, as well as, the challenges and future research directions are discussed in Section VII. Finally, Section VIII concludes the paper. 在本调查论文中，主要贡献与论文的组织结构相匹配。因此，本文的贡献/章节安排如下。第二节对现有的室内定位研究进行了比较。第三节介绍了 IPS 一般架构的定义，以及应急响应人员对 IPS 的主要要求。第四节全面介绍了用于测量距离、位移和航向的技术，以及定位方法及其利弊。第五节提出了一种新的分类方法，用于对用于应急响应人员的 IPS 方案进行分类。第六节对用于应急响应人员的最新 IPS 方案进行了全面调查，强调了每种方案的优缺点。本节最后对应急响应器 IPS 的最新技术进行了详细的比较总结，根据每种 IPS 的设计选择、要求和附加功能，强调了它们的异同（第 VI-E 节）。第 VII 部分讨论了与当前用于应急响应人员的 IPS 有关的问题，以及面临的挑战和未来的研究方向。最后，第八节为本文的结论。
II. Comparison of Survey Papers on Indoor Localization II.室内定位调查论文比较
TABLE I 表 I
Comparison of Survey Papers on Indoor Localization With Regard to Research Topics Reviewed and Comparison Criteria 关于室内定位的调查论文与所审查的研究课题和比较标准的比较
Survey 调查
Research Topics Reviewed 审查的研究课题
Comparison Criteria 比较标准
Year 年份
Localization in Wireless Sensor Networks (WSNs) 无线传感器网络（WSN）中的定位功能
Measurement techniques and localization algorithms (lateration, angulation, cellular proximity, dead reckoning, MLE, EKF, and PF) for MWSNs;
Centralized VS Distributed algorithms.
Localization in Mobile WSNs (MWSNs);
Measurement techniques and localization algorithms (lateration, angulation, cellular proximity, dead reckoning, MLE, EKF, and PF) for MWSNs;
Centralized VS Distributed algorithms.| Localization in Mobile WSNs (MWSNs); |
| :--- |
| Measurement techniques and localization algorithms (lateration, angulation, cellular proximity, dead reckoning, MLE, EKF, and PF) for MWSNs; |
| Centralized VS Distributed algorithms. |
No comparison criteria are proposed. 没有提出比较标准。
2009
Localization Techniques for WSNs [38] WSN 的定位技术 [38]
Localization techniques (ToA, TDoA, RToF, RSS, radio hop count, and AoA) and methods;
Comparison between centralized and distributed algorithms in WSNs.
Localization techniques (ToA, TDoA, RToF, RSS, radio hop count, and AoA) and methods;
Comparison between centralized and distributed algorithms in WSNs.| Localization techniques (ToA, TDoA, RToF, RSS, radio hop count, and AoA) and methods; |
| :--- |
| Comparison between centralized and distributed algorithms in WSNs. |
Centralized or distributed algorithms 集中式或分布式算法
2011
Localization algorithms of WSNs [44] WSN 的定位算法 [44]
Classification and comparison of localization algorithms based on the mobility state of landmarks and unknown nodes. 基于地标和未知节点移动状态的定位算法分类与比较。
Localization algorithms, accuracy, coverage, localization time, landmark number, and energy consumption. 定位算法、精度、覆盖范围、定位时间、地标数量和能耗。
2013
RSSI based Localization in WSNs [45] 基于 RSSI 的 WSN 定位 [45]
Models and techniques to reduce location errors and improve accuracy, for RSSI-based IPSs in WSNs. 减少 WSN 中基于 RSSI 的 IPS 的定位误差并提高准确性的模型和技术。
Context in use, methods, performance, and future scope. 使用背景、方法、性能和未来范围。
2015
Localization Techniques and Methods 定位技术和方法
无线 IPS 和技术
Wireless IPSs and
Techniques
Wireless IPSs and
Techniques| Wireless IPSs and |
| :--- |
| Techniques |
Detailed analysis on measuring principles and positioning algorithms, namely: Triangulation (ToA, TDoA, RSS, RToF, RSP, and AoA); Scene Analysis (fingerprinting); and Proximity;
A review of IPSs based on wireless technology: GPS-based RFID, cellularbased, UWB, WLAN, Bluetooth, and others (UHF multiple media, cordless phone system, and WSNs).
Detailed analysis on measuring principles and positioning algorithms, namely: Triangulation (ToA, TDoA, RSS, RToF, RSP, and AoA); Scene Analysis (fingerprinting); and Proximity;
A review of IPSs based on wireless technology: GPS-based RFID, cellularbased, UWB, WLAN, Bluetooth, and others (UHF multiple media, cordless phone system, and WSNs).| Detailed analysis on measuring principles and positioning algorithms, namely: Triangulation (ToA, TDoA, RSS, RToF, RSP, and AoA); Scene Analysis (fingerprinting); and Proximity; |
| :--- |
| A review of IPSs based on wireless technology: GPS-based RFID, cellularbased, UWB, WLAN, Bluetooth, and others (UHF multiple media, cordless phone system, and WSNs). |
Review of measurement techniques (ToA, TDoA, RToF, AoA, and RSS) and localization methods (trilateration, triangulation, and MLE);
Review of IPSs based on wireless technology (infrared, RF, and ultrasound).
Review of measurement techniques (ToA, TDoA, RToF, AoA, and RSS) and localization methods (trilateration, triangulation, and MLE);
Review of IPSs based on wireless technology (infrared, RF, and ultrasound).| Review of measurement techniques (ToA, TDoA, RToF, AoA, and RSS) and localization methods (trilateration, triangulation, and MLE); |
| :--- |
| Review of IPSs based on wireless technology (infrared, RF, and ultrasound). |
Wireless technology, accuracy, localization method, and overall evaluation. 无线技术、精确度、定位方法和总体评价。
2010
Inertial sensor-based methods in walking speed estimation [40] 基于惯性传感器的步行速度估算方法 [40]
Analysis of inertial sensor based IPSs for walking speed estimation. The existing methods are categorized by sensor specification, sensor attachment location, experimental design, and walking speed estimation algorithm. 分析基于惯性传感器的步行速度估算 IPS。现有方法按传感器规格、传感器安装位置、实验设计和步行速度估算算法进行了分类。
A review and comparison of techniques for step detection, characterization, inertial navigation, and step-heading-based dead reckoning;
Inertial navigation based on smartphones and hybrid systems.
A review and comparison of techniques for step detection, characterization, inertial navigation, and step-heading-based dead reckoning;
Inertial navigation based on smartphones and hybrid systems.| A review and comparison of techniques for step detection, characterization, inertial navigation, and step-heading-based dead reckoning; |
| :--- |
| Inertial navigation based on smartphones and hybrid systems. |
Step detection systems: sites, sensors, stance, signals used, and technique summary; Step and heading systems: system description, number of users, evaluation, particle scheme, number of initialization particle, number of tracking particles, initialization, and accuracy quoted. 台阶检测系统：地点、传感器、姿态、使用的信号和技术摘要；台阶和航向系统：系统说明、用户数量、评估、粒子方案、初始化粒子数、跟踪粒子数、初始化和引用的精度。
2013
Calibrationfree IPSs [47] 无校准 IPS [47]
Traditional and calibration-free performance comparison criteria; A review of calibration-free IPSs, based on the fingerprinting method. 传统和免校准性能比较标准；基于指纹法的免校准 IPS 评述。
Calibration-free: map requirement, acquiring location fix, seamless user participation, usage of indoor landmarks, need for additional sensors, and addressing device heterogeneity.
Traditional: accuracy, precision, scalability, robustness, security, privacy, cost, complexity, technology, and latency;
Calibration-free: map requirement, acquiring location fix, seamless user participation, usage of indoor landmarks, need for additional sensors, and addressing device heterogeneity.| Traditional: accuracy, precision, scalability, robustness, security, privacy, cost, complexity, technology, and latency; |
| :--- |
| Calibration-free: map requirement, acquiring location fix, seamless user participation, usage of indoor landmarks, need for additional sensors, and addressing device heterogeneity. |
A review of measurements types (geometric-related measurements, position-related measurements, and self-measurements), methods (Bayesian tracking, distributed and cooperative tracking, fingerprinting, SLAM, fusion, and others), and technologies (short-range wireless technologies, cellular networks, UWB, and near-field) for indoor tracking 综述用于室内跟踪的测量类型（几何相关测量、位置相关测量和自测量）、方法（贝叶斯跟踪、分布式合作跟踪、指纹识别、SLAM、融合等）和技术（短程无线技术、蜂窝网络、UWB 和近场）。
Technology, measurement technique, accuracy, pros, and cons. 技术、测量技术、精度、优点和缺点。
2015
Wi-Fi 指纹技术的最新进展 [1]
Recent
Advances on
Wi-Fi
Fingerprintin
g [1]
Recent
Advances on
Wi-Fi
Fingerprintin
g [1]| Recent |
| :--- |
| Advances on |
| Wi-Fi |
| Fingerprintin |
| g [1] |
Advanced localization techniques based on the fingerprinting method. Exploiting the use of spatial and temporal signal patterns, collaborative localization, and motion-assisted localization;
Efficient system deployment. With the focus on reducing offline site survey, adapting to fingerprint changes, addressing device heterogeneity, and reducing the energy consumption of Wi-Fi fingerprint method.
Advanced localization techniques based on the fingerprinting method. Exploiting the use of spatial and temporal signal patterns, collaborative localization, and motion-assisted localization;
Efficient system deployment. With the focus on reducing offline site survey, adapting to fingerprint changes, addressing device heterogeneity, and reducing the energy consumption of Wi-Fi fingerprint method.| Advanced localization techniques based on the fingerprinting method. Exploiting the use of spatial and temporal signal patterns, collaborative localization, and motion-assisted localization; |
| :--- |
| Efficient system deployment. With the focus on reducing offline site survey, adapting to fingerprint changes, addressing device heterogeneity, and reducing the energy consumption of Wi-Fi fingerprint method. |
Localization technology, service features, other features, reported mean accuracy, and year of establishment. 定位技术、服务功能、其他功能、报告的平均精确度以及成立年份。
2016
Survey Research Topics Reviewed Comparison Criteria Year
Localization in Wireless Sensor Networks (WSNs)
WSN localization techniques [32] "Measurement techniques (AoA, ToA, RToF, Lighthouse approach, TDoA, RSS) and one-hop localization algorithms (Triangulation, Trilateration, Fingerprinting, Lighthouse approach, Data Fusion of hybrid measurements); Connectivity-based multi-hop localization algorithms;
Distance-based multi-hop localization algorithms. Centralized VS Distributed algorithms." No comparison criteria are proposed. 2007
Localization for Mobile WSN [34] "Localization in Mobile WSNs (MWSNs);
Measurement techniques and localization algorithms (lateration, angulation, cellular proximity, dead reckoning, MLE, EKF, and PF) for MWSNs;
Centralized VS Distributed algorithms." No comparison criteria are proposed. 2009
Localization Techniques for WSNs [38] "Localization techniques (ToA, TDoA, RToF, RSS, radio hop count, and AoA) and methods;
Comparison between centralized and distributed algorithms in WSNs." Centralized or distributed algorithms 2011
Localization algorithms of WSNs [44] Classification and comparison of localization algorithms based on the mobility state of landmarks and unknown nodes. Localization algorithms, accuracy, coverage, localization time, landmark number, and energy consumption. 2013
RSSI based Localization in WSNs [45] Models and techniques to reduce location errors and improve accuracy, for RSSI-based IPSs in WSNs. Context in use, methods, performance, and future scope. 2015
Localization Techniques and Methods
"Wireless IPSs and
Techniques" "Detailed analysis on measuring principles and positioning algorithms, namely: Triangulation (ToA, TDoA, RSS, RToF, RSP, and AoA); Scene Analysis (fingerprinting); and Proximity;
A review of IPSs based on wireless technology: GPS-based RFID, cellularbased, UWB, WLAN, Bluetooth, and others (UHF multiple media, cordless phone system, and WSNs)." Accuracy, precision, complexity, robustness, scalability, cost; wireless technologies, and positioning algorithm. 2007
Technologies for Indoor Human Tracking [35] "Review of measurement techniques (ToA, TDoA, RToF, AoA, and RSS) and localization methods (trilateration, triangulation, and MLE);
Review of IPSs based on wireless technology (infrared, RF, and ultrasound)." Wireless technology, accuracy, localization method, and overall evaluation. 2010
Inertial sensor-based methods in walking speed estimation [40] Analysis of inertial sensor based IPSs for walking speed estimation. The existing methods are categorized by sensor specification, sensor attachment location, experimental design, and walking speed estimation algorithm. No comparison criteria are proposed. 2012
Inertial IPSs for pedestrians [42] "A review and comparison of techniques for step detection, characterization, inertial navigation, and step-heading-based dead reckoning;
Inertial navigation based on smartphones and hybrid systems." Step detection systems: sites, sensors, stance, signals used, and technique summary; Step and heading systems: system description, number of users, evaluation, particle scheme, number of initialization particle, number of tracking particles, initialization, and accuracy quoted. 2013
Calibrationfree IPSs [47] Traditional and calibration-free performance comparison criteria; A review of calibration-free IPSs, based on the fingerprinting method. "Traditional: accuracy, precision, scalability, robustness, security, privacy, cost, complexity, technology, and latency;
Calibration-free: map requirement, acquiring location fix, seamless user participation, usage of indoor landmarks, need for additional sensors, and addressing device heterogeneity." 2015
" Indoor
Tracking [48] " A review of measurements types (geometric-related measurements, position-related measurements, and self-measurements), methods (Bayesian tracking, distributed and cooperative tracking, fingerprinting, SLAM, fusion, and others), and technologies (short-range wireless technologies, cellular networks, UWB, and near-field) for indoor tracking Technology, measurement technique, accuracy, pros, and cons. 2015
"Recent
Advances on
Wi-Fi
Fingerprintin
g [1]" "Advanced localization techniques based on the fingerprinting method. Exploiting the use of spatial and temporal signal patterns, collaborative localization, and motion-assisted localization;
Efficient system deployment. With the focus on reducing offline site survey, adapting to fingerprint changes, addressing device heterogeneity, and reducing the energy consumption of Wi-Fi fingerprint method." Localization technology, service features, other features, reported mean accuracy, and year of establishment. 2016| Survey | Research Topics Reviewed | Comparison Criteria | Year |
| :---: | :---: | :---: | :---: |
| Localization in Wireless Sensor Networks (WSNs) | | | |
| WSN localization techniques [32] | Measurement techniques (AoA, ToA, RToF, Lighthouse approach, TDoA, RSS) and one-hop localization algorithms (Triangulation, Trilateration, Fingerprinting, Lighthouse approach, Data Fusion of hybrid measurements); Connectivity-based multi-hop localization algorithms; <br> Distance-based multi-hop localization algorithms. Centralized VS Distributed algorithms. | No comparison criteria are proposed. | 2007 |
| Localization for Mobile WSN [34] | Localization in Mobile WSNs (MWSNs); <br> Measurement techniques and localization algorithms (lateration, angulation, cellular proximity, dead reckoning, MLE, EKF, and PF) for MWSNs; <br> Centralized VS Distributed algorithms. | No comparison criteria are proposed. | 2009 |
| Localization Techniques for WSNs [38] | Localization techniques (ToA, TDoA, RToF, RSS, radio hop count, and AoA) and methods; <br> Comparison between centralized and distributed algorithms in WSNs. | Centralized or distributed algorithms | 2011 |
| Localization algorithms of WSNs [44] | Classification and comparison of localization algorithms based on the mobility state of landmarks and unknown nodes. | Localization algorithms, accuracy, coverage, localization time, landmark number, and energy consumption. | 2013 |
| RSSI based Localization in WSNs [45] | Models and techniques to reduce location errors and improve accuracy, for RSSI-based IPSs in WSNs. | Context in use, methods, performance, and future scope. | 2015 |
| Localization Techniques and Methods | | | |
| Wireless IPSs and <br> Techniques | Detailed analysis on measuring principles and positioning algorithms, namely: Triangulation (ToA, TDoA, RSS, RToF, RSP, and AoA); Scene Analysis (fingerprinting); and Proximity; <br> A review of IPSs based on wireless technology: GPS-based RFID, cellularbased, UWB, WLAN, Bluetooth, and others (UHF multiple media, cordless phone system, and WSNs). | Accuracy, precision, complexity, robustness, scalability, cost; wireless technologies, and positioning algorithm. | 2007 |
| Technologies for Indoor Human Tracking [35] | Review of measurement techniques (ToA, TDoA, RToF, AoA, and RSS) and localization methods (trilateration, triangulation, and MLE); <br> Review of IPSs based on wireless technology (infrared, RF, and ultrasound). | Wireless technology, accuracy, localization method, and overall evaluation. | 2010 |
| Inertial sensor-based methods in walking speed estimation [40] | Analysis of inertial sensor based IPSs for walking speed estimation. The existing methods are categorized by sensor specification, sensor attachment location, experimental design, and walking speed estimation algorithm. | No comparison criteria are proposed. | 2012 |
| Inertial IPSs for pedestrians [42] | A review and comparison of techniques for step detection, characterization, inertial navigation, and step-heading-based dead reckoning; <br> Inertial navigation based on smartphones and hybrid systems. | Step detection systems: sites, sensors, stance, signals used, and technique summary; Step and heading systems: system description, number of users, evaluation, particle scheme, number of initialization particle, number of tracking particles, initialization, and accuracy quoted. | 2013 |
| Calibrationfree IPSs [47] | Traditional and calibration-free performance comparison criteria; A review of calibration-free IPSs, based on the fingerprinting method. | Traditional: accuracy, precision, scalability, robustness, security, privacy, cost, complexity, technology, and latency; <br> Calibration-free: map requirement, acquiring location fix, seamless user participation, usage of indoor landmarks, need for additional sensors, and addressing device heterogeneity. | 2015 |
| $\begin{aligned} & \text { Indoor } \\ & \text { Tracking [48] } \end{aligned}$ | A review of measurements types (geometric-related measurements, position-related measurements, and self-measurements), methods (Bayesian tracking, distributed and cooperative tracking, fingerprinting, SLAM, fusion, and others), and technologies (short-range wireless technologies, cellular networks, UWB, and near-field) for indoor tracking | Technology, measurement technique, accuracy, pros, and cons. | 2015 |
| Recent <br> Advances on <br> Wi-Fi <br> Fingerprintin <br> g [1] | Advanced localization techniques based on the fingerprinting method. Exploiting the use of spatial and temporal signal patterns, collaborative localization, and motion-assisted localization; <br> Efficient system deployment. With the focus on reducing offline site survey, adapting to fingerprint changes, addressing device heterogeneity, and reducing the energy consumption of Wi-Fi fingerprint method. | Localization technology, service features, other features, reported mean accuracy, and year of establishment. | 2016 |
(Continued on the next page) (接下页）
Several works have been conducted focusing on localization in WSNs [32], [34], [38], [44], [45]. This indoor localization problem has been attracting several researchers due to the role that localization plays within WSNs. Mao et al. [32] 目前已有多项研究集中在 WSN 的定位问题上 [32], [34], [38], [44], [45]。由于定位在 WSN 中的作用，这一室内定位问题一直吸引着一些研究人员。Mao 等人[32］
conducted a comprehensive review on the measurement techniques applied within WSNs including a deep analysis of the localization algorithms for WSNs and their classification as one hop, connectivity-based multi-hop, and 对应用于 WSN 的测量技术进行了全面回顾，包括深入分析 WSN 的定位算法及其一跳、基于连接的多跳和多跳分类。
TABLE I 表 I
To Be Continued 待续
Survey 调查
Research Topics Reviewed 审查的研究课题
Comparison Criteria 比较标准
Year 年份
Localization Technologies 本地化技术
IPSs for Wireless Personal Networks [33] 无线个人网络的 IPS [33]
A review of IPSs for PNs based on the wireless technology (Infrared, Ultrasound, RF, Magnetic, Vision-based, and Audible Sound);
A brief description of techniques and methods for localization, with the focus on methods based on geometric proprieties (ToA, AoA, RSS), fingerprinting, proximity and vision analysis.
A review of IPSs for PNs based on the wireless technology (Infrared, Ultrasound, RF, Magnetic, Vision-based, and Audible Sound);
A brief description of techniques and methods for localization, with the focus on methods based on geometric proprieties (ToA, AoA, RSS), fingerprinting, proximity and vision analysis.| A review of IPSs for PNs based on the wireless technology (Infrared, Ultrasound, RF, Magnetic, Vision-based, and Audible Sound); |
| :--- |
| A brief description of techniques and methods for localization, with the focus on methods based on geometric proprieties (ToA, AoA, RSS), fingerprinting, proximity and vision analysis. |
Security and privacy, cost; performance, robustness and fault tolerance, complexity, user preference, commercial availability, and limitations. 安全和隐私、成本、性能、稳健性和容错性、复杂性、用户偏好、商业可用性和局限性。
2009
Localization Technologies Towards Navigational [36] 面向导航的定位技术 [36]
Reviews the localization process based on the dominant technologies (RF, photonic, sonic, inertial sensors, atmospheric pressure, and magnetic); 回顾基于主流技术（射频、光子、声波、惯性传感器、大气压力和磁性）的定位过程；
Main technology, technology details, localization technique, disadvantages, advantages, position, and orientation. 主要技术、技术细节、定位技术、缺点、优点、位置和方向。
2010
Optical-based IPSs [37] 基于光学的 IPS [37]
A review and comparison of optical IPS based on how the images are referenced to the environment (reference from 3D building models, reference from images, reference from deployed coded targets, reference from projected targets, systems without reference, and reference from other sensors). 根据图像与环境的参照方式（三维建筑模型参照、图像参照、部署的编码目标参照、投射目标参照、无参照系统以及其他传感器参照）对光学 IPS 进行审查和比较。
Coordinate reference, reported accuracy, coverage, CCD size, frame rate, object/camera positioning, camera cost, and market maturity. 坐标参考、报告精度、覆盖范围、CCD 尺寸、帧频、物体/摄像机定位、摄像机成本和市场成熟度。
Classification of IPSs based on the use of active or passive localization techniques;
Within the passive or active group, each IPS is classified based on localization technology.
Classification of IPSs based on the use of active or passive localization techniques;
Within the passive or active group, each IPS is classified based on localization technology.| Classification of IPSs based on the use of active or passive localization techniques; |
| :--- |
| Within the passive or active group, each IPS is classified based on localization technology. |
Localization techniques (AoA, ToA, TDoA, RToF, and RSSI) and methods (proximity, triangulation, trilateration, dead reckoning, and map matching);
A brief review of wireless IPSs based on GPS, infrared radiation, RF (RFID, Bluetooth, UWB, FM radio, and ZigBee), ultrasound, and hybrid IPSs;
A detailed review of wireless IPSs based on WLAN technology (Wi-Fi-based and fingerprinting based).
Localization techniques (AoA, ToA, TDoA, RToF, and RSSI) and methods (proximity, triangulation, trilateration, dead reckoning, and map matching);
A brief review of wireless IPSs based on GPS, infrared radiation, RF (RFID, Bluetooth, UWB, FM radio, and ZigBee), ultrasound, and hybrid IPSs;
A detailed review of wireless IPSs based on WLAN technology (Wi-Fi-based and fingerprinting based).| Localization techniques (AoA, ToA, TDoA, RToF, and RSSI) and methods (proximity, triangulation, trilateration, dead reckoning, and map matching); |
| :--- |
| A brief review of wireless IPSs based on GPS, infrared radiation, RF (RFID, Bluetooth, UWB, FM radio, and ZigBee), ultrasound, and hybrid IPSs; |
| A detailed review of wireless IPSs based on WLAN technology (Wi-Fi-based and fingerprinting based). |
Accuracy, principles used for localization, coverage, power consumption, cost and remarks. 精度、定位原理、覆盖范围、功耗、成本和备注。
2013
Mobility Increases Localizability on Wireless IPSs using IMUs [46] 移动性提高了使用 IMU 的无线 IPS 的可定位性 [46]
How to measure human mobility, what type of sensors can be used and what types of mobility information can be acquired;
Handling measurement errors of smartphone built-in sensors.
Mobility enhancing smartphone-based IPS;
How to measure human mobility, what type of sensors can be used and what types of mobility information can be acquired;
Handling measurement errors of smartphone built-in sensors.| Mobility enhancing smartphone-based IPS; |
| :--- |
| How to measure human mobility, what type of sensors can be used and what types of mobility information can be acquired; |
| Handling measurement errors of smartphone built-in sensors. |
Sensor, particle model, feature, and accuracy. 传感器、粒子模型、特征和精度。
2015
IPSs for Emergency Responders 应急响应人员的 IPS
Localization Support for Emergency Responders [22] 为应急响应人员提供本地化支持 [22]
综述用于应急响应人员的 IPS。未提出分类法；简要介绍用于定位的方法（基于基础设施的定位、利用 WSN 的定位、Ad Hoc 相对定位、近距离定位和死位推算）。
A review of IPSs for emergency responders. No taxonomy is proposed;
A brief description of the methods used for localization (infrastructure-based localization, localization with WSNs, Ad Hoc relative positioning, proximity, and dead reckoning).
A review of IPSs for emergency responders. No taxonomy is proposed;
A brief description of the methods used for localization (infrastructure-based localization, localization with WSNs, Ad Hoc relative positioning, proximity, and dead reckoning).| A review of IPSs for emergency responders. No taxonomy is proposed; |
| :--- |
| A brief description of the methods used for localization (infrastructure-based localization, localization with WSNs, Ad Hoc relative positioning, proximity, and dead reckoning). |
Primary function, amount of information and flexibility, technology, components, deployment and prior knowledge, limitation, and additional features. 主要功能、信息量和灵活性、技术、组件、部署和先前知识、限制和附加功能。
2010
Indoor tracking for mission critical scenarios [12] 关键任务场景的室内跟踪 [12]
A brief description of requirements within mission critical scenarios;
A brief review of localization methods (angulation, lateration, fingerprinting, inertial sensors, and connectivity);
Analysis of existing IPSs for emergency responders based on techniques employed for localization (signal related or IMU sensing), with the focus on tracking capabilities.
A brief description of requirements within mission critical scenarios;
A brief review of localization methods (angulation, lateration, fingerprinting, inertial sensors, and connectivity);
Analysis of existing IPSs for emergency responders based on techniques employed for localization (signal related or IMU sensing), with the focus on tracking capabilities.| A brief description of requirements within mission critical scenarios; |
| :--- |
| A brief review of localization methods (angulation, lateration, fingerprinting, inertial sensors, and connectivity); |
| Analysis of existing IPSs for emergency responders based on techniques employed for localization (signal related or IMU sensing), with the focus on tracking capabilities. |
Survey Research Topics Reviewed Comparison Criteria Year
Localization Technologies
IPSs for Wireless Personal Networks [33] "A review of IPSs for PNs based on the wireless technology (Infrared, Ultrasound, RF, Magnetic, Vision-based, and Audible Sound);
A brief description of techniques and methods for localization, with the focus on methods based on geometric proprieties (ToA, AoA, RSS), fingerprinting, proximity and vision analysis." Security and privacy, cost; performance, robustness and fault tolerance, complexity, user preference, commercial availability, and limitations. 2009
Localization Technologies Towards Navigational [36] Reviews the localization process based on the dominant technologies (RF, photonic, sonic, inertial sensors, atmospheric pressure, and magnetic); Main technology, technology details, localization technique, disadvantages, advantages, position, and orientation. 2010
Optical-based IPSs [37] A review and comparison of optical IPS based on how the images are referenced to the environment (reference from 3D building models, reference from images, reference from deployed coded targets, reference from projected targets, systems without reference, and reference from other sensors). Coordinate reference, reported accuracy, coverage, CCD size, frame rate, object/camera positioning, camera cost, and market maturity. 2011
Active and Passive IPSs [39] "Classification of IPSs based on the use of active or passive localization techniques;
Within the passive or active group, each IPS is classified based on localization technology." Wireless technology, positioning algorithm, accuracy/precision, complexity, scalability/space dimension, and cost. 2012
Ultrasonic-based IPSs [41] A detailed review of ultrasonic-based IPSs. Spreading and channel access, update rate, measurement method, accuracy, orientation, structure, and cost. 2013
Recent Advances in Wireless IPSs and Techniques [43] "Localization techniques (AoA, ToA, TDoA, RToF, and RSSI) and methods (proximity, triangulation, trilateration, dead reckoning, and map matching);
A brief review of wireless IPSs based on GPS, infrared radiation, RF (RFID, Bluetooth, UWB, FM radio, and ZigBee), ultrasound, and hybrid IPSs;
A detailed review of wireless IPSs based on WLAN technology (Wi-Fi-based and fingerprinting based)." Accuracy, principles used for localization, coverage, power consumption, cost and remarks. 2013
Mobility Increases Localizability on Wireless IPSs using IMUs [46] "Mobility enhancing smartphone-based IPS;
How to measure human mobility, what type of sensors can be used and what types of mobility information can be acquired;
Handling measurement errors of smartphone built-in sensors." Sensor, particle model, feature, and accuracy. 2015
IPSs for Emergency Responders
Localization Support for Emergency Responders [22] "A review of IPSs for emergency responders. No taxonomy is proposed;
A brief description of the methods used for localization (infrastructure-based localization, localization with WSNs, Ad Hoc relative positioning, proximity, and dead reckoning)." Primary function, amount of information and flexibility, technology, components, deployment and prior knowledge, limitation, and additional features. 2010
Indoor tracking for mission critical scenarios [12] "A brief description of requirements within mission critical scenarios;
A brief review of localization methods (angulation, lateration, fingerprinting, inertial sensors, and connectivity);
Analysis of existing IPSs for emergency responders based on techniques employed for localization (signal related or IMU sensing), with the focus on tracking capabilities." Sensors, localization method, precision, deployability, complexity, cost, remarks 2011| Survey | Research Topics Reviewed | Comparison Criteria | Year |
| :---: | :---: | :---: | :---: |
| Localization Technologies | | | |
| IPSs for Wireless Personal Networks [33] | A review of IPSs for PNs based on the wireless technology (Infrared, Ultrasound, RF, Magnetic, Vision-based, and Audible Sound); <br> A brief description of techniques and methods for localization, with the focus on methods based on geometric proprieties (ToA, AoA, RSS), fingerprinting, proximity and vision analysis. | Security and privacy, cost; performance, robustness and fault tolerance, complexity, user preference, commercial availability, and limitations. | 2009 |
| Localization Technologies Towards Navigational [36] | Reviews the localization process based on the dominant technologies (RF, photonic, sonic, inertial sensors, atmospheric pressure, and magnetic); | Main technology, technology details, localization technique, disadvantages, advantages, position, and orientation. | 2010 |
| Optical-based IPSs [37] | A review and comparison of optical IPS based on how the images are referenced to the environment (reference from 3D building models, reference from images, reference from deployed coded targets, reference from projected targets, systems without reference, and reference from other sensors). | Coordinate reference, reported accuracy, coverage, CCD size, frame rate, object/camera positioning, camera cost, and market maturity. | 2011 |
| Active and Passive IPSs [39] | Classification of IPSs based on the use of active or passive localization techniques; <br> Within the passive or active group, each IPS is classified based on localization technology. | Wireless technology, positioning algorithm, accuracy/precision, complexity, scalability/space dimension, and cost. | 2012 |
| Ultrasonic-based IPSs [41] | A detailed review of ultrasonic-based IPSs. | Spreading and channel access, update rate, measurement method, accuracy, orientation, structure, and cost. | 2013 |
| Recent Advances in Wireless IPSs and Techniques [43] | Localization techniques (AoA, ToA, TDoA, RToF, and RSSI) and methods (proximity, triangulation, trilateration, dead reckoning, and map matching); <br> A brief review of wireless IPSs based on GPS, infrared radiation, RF (RFID, Bluetooth, UWB, FM radio, and ZigBee), ultrasound, and hybrid IPSs; <br> A detailed review of wireless IPSs based on WLAN technology (Wi-Fi-based and fingerprinting based). | Accuracy, principles used for localization, coverage, power consumption, cost and remarks. | 2013 |
| Mobility Increases Localizability on Wireless IPSs using IMUs [46] | Mobility enhancing smartphone-based IPS; <br> How to measure human mobility, what type of sensors can be used and what types of mobility information can be acquired; <br> Handling measurement errors of smartphone built-in sensors. | Sensor, particle model, feature, and accuracy. | 2015 |
| IPSs for Emergency Responders | | | |
| Localization Support for Emergency Responders [22] | A review of IPSs for emergency responders. No taxonomy is proposed; <br> A brief description of the methods used for localization (infrastructure-based localization, localization with WSNs, Ad Hoc relative positioning, proximity, and dead reckoning). | Primary function, amount of information and flexibility, technology, components, deployment and prior knowledge, limitation, and additional features. | 2010 |
| Indoor tracking for mission critical scenarios [12] | A brief description of requirements within mission critical scenarios; <br> A brief review of localization methods (angulation, lateration, fingerprinting, inertial sensors, and connectivity); <br> Analysis of existing IPSs for emergency responders based on techniques employed for localization (signal related or IMU sensing), with the focus on tracking capabilities. | Sensors, localization method, precision, deployability, complexity, cost, remarks | 2011 |
Measurement techniques and localization methods are other topics of indoor localization that have been analyzed in several papers [1], [31], [35], [40], [42], [47], [48]. Liu et al. [31] conducted a detailed survey on wireless measurement techniques and localization methods together with a comprehensive review of wireless IPSs, highlighting the pros and 测量技术和定位方法是室内定位的其他主题，已有多篇论文[1]、[31]、[35]、[40]、[42]、[47]、[48]对其进行了分析。Liu 等人[31] 对无线测量技术和定位方法进行了详细调查，并对无线 IPS 进行了全面评述，重点介绍了其优点和缺点。
cons of each wireless technology used. Nevertheless, the Inertial Measurement Unit (IMU)-based solutions are not addressed in the paper. Similar work was carried out by Zhang et al. [35], but the discussion on existing IPSs is not sufficiently deep. The surveys from Yang and Li [40] and from Harle [42] are focused on IPSs based on inertial sensors. The former concentrates on describing methods for IMUbased walking speed estimation, while the latter provides a detailed analysis of the different subsystems of an IMUbased IPS (step detection, characterization, inertial navigation, and step-heading dead reckoning), further comparing the IPSs for the step detection and the step-heading subsystems. Dardari et al. [48] focused their survey on the measurement types, methods, and technologies for indoor tracking. Due to the attractiveness of fingerprinting-based IPSs, several works have been proposed to decrease the deployment effort and increase the accuracy of the IPS. In the survey papers of Hossain and Soh [47] and He and Chan [1] the emerging fingerprinting solutions are reviewed. In both reviews, the solutions proposed to relieve pre-deployment woes are addressed. In addition, the work from He and Chan [1] reviews the advanced localization techniques, namely on fingerprintingbased algorithms that exploit spatial and temporal patterns, collaborative localization, and motion-assisted localization. On the other hand, the major contribution of the work conducted by Hossain and Soh [47] is the differentiation between traditional and calibration-free comparison criteria. Although these works are impressive, none of them address the specific scenarios and constraints that emergency responders face during their missions. 每种无线技术的缺点。不过，本文并未涉及基于惯性测量单元（IMU）的解决方案。Zhang 等人[35]也开展了类似的工作，但对现有 IPS 的讨论不够深入。Yang和Li[40]以及Harle[42]的研究主要集中在基于惯性传感器的IPS上。前者集中介绍了基于惯性传感器的步行速度估算方法，后者则详细分析了基于惯性传感器的 IPS 的不同子系统（台阶检测、特征描述、惯性导航和台阶定向推算），并进一步比较了台阶检测和台阶定向子系统的 IPS。Dardari 等人[48]重点调查了室内跟踪的测量类型、方法和技术。由于基于指纹识别的 IPS 很有吸引力，人们提出了几项工作来减少部署工作量并提高 IPS 的准确性。在 Hossain 和 Soh [47] 以及 He 和 Chan [1] 的调查论文中，对新出现的指纹识别解决方案进行了综述。在这两篇综述中，都讨论了为缓解部署前的困境而提出的解决方案。此外，He 和 Chan [1] 的研究还回顾了先进的定位技术，即利用空间和时间模式的基于指纹的算法、协作定位和运动辅助定位。另一方面，Hossain 和 Soh [47] 所做工作的主要贡献在于区分了传统比较标准和免校准比较标准。尽管这些工作令人印象深刻，但它们都没有解决应急响应人员在执行任务时面临的特定场景和限制。
Another focus of survey papers on indoor localization is based on the classification of IPSs according to the technology employed [33], [36], [37], [39], [41], [43], [46]. A single technology review was carried out by Mautz and Tilch [37] and by Ijaz et al. [41], the former focusing on opticalbased IPSs and the latter on ultrasonic-based IPSs. Survey papers like [29], [32], and [39] classified the IPSs based on technology used. Gu et al. [33] reviewed the wireless IPSs and emphasized the pros and cons of each technology analyzed. IMU-based approaches are not addressed. The wireless- and IMU-based approaches were surveyed on the work of Torre-Solis et al. [36], comparing the IPSs on the basis of navigational assistance for topographical disorientation. Farid et al. [43] review covered a vast range of technologies, but only detailing the IPSs based on WLAN technology. A survey paper comparing the active and passive IPSs was presented by Deak et al. [39]. Finally, Yang et al. [46] focused the review on a recent trend in indoor localization based on the use of a smartphone to compute the user’s location. Again, the works here presented are impressive and cover a vast range of technologies, but they do not address the challenges faced by emergency responders in an emergency scenario. 室内定位调查论文的另一个重点是根据所采用的技术对 IPS 进行分类 [33]、[36]、[37]、[39]、[41]、[43]、[46]。Mautz 和 Tilch[37]以及 Ijaz 等人[41]对单一技术进行了综述，前者侧重于基于光学的 IPS，后者侧重于基于超声波的 IPS。文献[29]、[32]和[39]等调查论文根据所使用的技术对 IPS 进行了分类。Gu 等人[33] 综述了无线 IPS，并强调了所分析的每种技术的优缺点。基于 IMU 的方法没有涉及。Torre-Solis 等人[36]对基于无线和 IMU 的方法进行了调查，比较了基于地形迷失导航辅助的 IPS。Farid 等人[43] 的综述涵盖了大量技术，但只详细介绍了基于无线局域网技术的 IPS。Deak 等人[39]发表了一篇调查论文，对主动式和被动式 IPS 进行了比较。最后，Yang 等人[46] 的综述重点是基于使用智能手机计算用户位置的室内定位的最新趋势。同样，这里介绍的作品令人印象深刻，涵盖了大量技术，但它们并没有解决应急响应人员在紧急情况下面临的挑战。
From the analysis performed by the various surveys on indoor localization, only two were found addressing the challenging research area of IPSs for emergency responders [12], [22]. A description of different localization strategies for emergency responders’ IPSs, as well as, of some 在对各种室内定位调查进行的分析中，仅发现两项调查涉及应急响应人员 IPS 这一具有挑战性的研究领域[12]、[22]。对应急响应人员 IPS 不同定位策略的描述，以及对一些应急响应人员 IPS 的定位策略的描述，将有助于我们更好地了解应急响应人员 IPS。
Fig. 1. Illustration of the localization process. Representation of the two possible phases of the localization process, offline and online. With an inset on the signal measurement and position calculation phases of the online phase. 图 1.本地化过程示意图。定位过程的两个可能阶段：离线和在线。插图显示了在线阶段的信号测量和位置计算阶段。
IPSs developed for emergency responders is presented by Fischer and Gellersen [22]. Nevertheless, they do not propose a taxonomy and the description of both localization strategies and IPSs is presented at a very high level. Fuchs et al. [12] review classifies the existing IPSs for emergency responders as signal-based or IMU-based systems. A succinct description of user requirements and localization methods is presented. Hybrid IPSs are not addressed. Both survey papers refer to IPSs for emergency responders’ works until 2011. Fischer 和 Gellersen [22]介绍了为应急响应人员开发的 IPS。不过，他们并没有提出分类方法，对定位策略和 IPS 的描述都是在非常高的层次上进行的。Fuchs 等人[12]对现有的用于应急响应的 IPS 进行了分类，分为基于信号的系统和基于 IMU 的系统。文中简要介绍了用户需求和定位方法。混合型 IPS 未作讨论。这两篇调查论文都提到了 2011 年之前用于应急响应人员的 IPS。
III. IPS OvERVIEw and Requirements for Emergency Responders III.IPS OvERVIEw 和对应急响应人员的要求
By definition, an IPS is a system that is capable of processing the position of an object/person in a confined and closed physical space (e.g., hospital, gym, university, or museum), in a continuous way and in real time [33], [35], and where other systems, like GPS-based, do not work. 根据定义，IPS 是一种能够在密闭的物理空间（如医院、体育馆、大学或博物馆）中连续、实时地处理物体/人员位置的系统[33]、[35]。
Generically, an IPS can be composed of two distinct phases of the localization process, the offline and the online phase (Fig. 1). The offline phase is typical of IPSs that rely on the fingerprinting or visual analysis methods, further investigated in the Section IV-B. During the offline phase, the system will gather useful information about the environment through an extensive and labor-intensive site surveying. During the process of site surveying, some unique information about the propagation of radio signals at one specific point (fingerprints) or unique features of the building structure (landmarks in optical systems) are acquired and stored in a database. The landmarks and fingerprints are then associated with the physical position where they were collected. The database containing this building specific information will be used during the online phase to assist the localization process. 一般来说，IPS 可由定位过程的两个不同阶段组成，即离线阶段和在线阶段（图 1）。离线阶段是依赖指纹识别或视觉分析方法的 IPS 的典型阶段，将在第四节 B 部分进一步研究。在离线阶段，系统将通过大量劳动密集型现场勘测来收集有关环境的有用信息。在现场勘测过程中，系统会获取有关无线电信号在某一特定点传播的独特信息（指纹）或建筑物结构的独特特征（光学系统中的地标），并将其存储到数据库中。然后将地标和指纹与采集它们的物理位置关联起来。包含这些建筑物特定信息的数据库将在联机阶段用于辅助定位过程。
During the online phase, the IPS will compute the position of the user/object. This phase can rely or not on the offline phase, depending on the scheme defined for the IPS by the designers. The positioning process of the online phase 在在线阶段，IPS 将计算用户/物体的位置。这一阶段可以依赖或不依赖离线阶段，具体取决于设计人员为 IPS 定义的方案。在线阶段的定位过程
Fig. 2. Taxonomy of IPS requirements for emergency responders. 图 2.应急响应人员对 IPS 需求的分类。
can be further divided into two different phases: 1) the signal measurement, and 2) the position calculation (Fig. 1). The signal measurement phase is characterized by the collection of specific parameters through the use of proprioceptive and/or exteroceptive sensors. The exteroceptive sensors are sensors that react to external stimuli. Examples of exteroceptive sensors are the wireless receivers of the technologies chosen for the IPS (e.g., UWB, Wi-Fi, ZigBee, ultrasound), or sensors like barometers. Whereas, the proprioceptive sensors are sensors that react to the stimuli produced by the person/object to which they are attached. This class of sensors is usually associated with the change in position and movement of the body. Inertial sensors like accelerometers, gyroscopes, and magnetometers are examples of proprioceptive sensors. Both classes of sensors are used to measure the target physical quantities of a phenomenon at each timestamp. Also, they can be used separately, or they can be merged with data fusion algorithms. So, if the IPS is based on wireless technology, the system will exchange data between the mobile node, attached to the object/person, and the reference or anchor nodes, whose position is known a priori. Based on this data exchange, the system will collect some properties of the communication, e.g., ToA, TDoA, RSS, and AoA [35]. These and other distance measurement techniques are detailed in Section IV-A. 可进一步分为两个不同阶段：1) 信号测量，2) 位置计算（图 1）。信号测量阶段的特点是通过使用本体感觉和/或外部感觉传感器收集特定参数。外感觉传感器是对外部刺激做出反应的传感器。外感传感器的例子包括 IPS 所选技术的无线接收器（如 UWB、Wi-Fi、ZigBee、超声波）或气压计等传感器。而本体感觉传感器则是对所连接的人/物体产生的刺激做出反应的传感器。这类传感器通常与身体位置和运动的变化有关。加速度计、陀螺仪和磁力计等惯性传感器就是本体感觉传感器的例子。这两类传感器都用于测量每个时间戳的目标物理量。此外，它们可以单独使用，也可以与数据融合算法合并使用。因此，如果 IPS 基于无线技术，系统将在附着在物体/人员上的移动节点与位置先验已知的参考节点或锚节点之间交换数据。基于这种数据交换，系统将收集通信的一些属性，如 ToA、TDoA、RSS 和 AoA [35]。这些及其他距离测量技术详见第 IV-A 节。
During the position calculation phase, the IPS will compute the location of the mobile node, based on the data collected in the first phase. So, if the IPS is based on wireless communications technologies, geometric approaches like triangulation, trilateration, among others will be used to compute the position of the mobile unit. Additionally, optimization techniques 在位置计算阶段，IPS 将根据第一阶段收集的数据计算移动节点的位置。因此，如果 IPS 基于无线通信技术，则将使用三角测量、三坐标等几何方法来计算移动装置的位置。此外，优化技术
based on statistical models can be employed to filter out the noise and thus increase the system’s accuracy [35], [48]. This noise in the measurements is the result of countless sources of interferences that a signal faces in real indoor environments. On the other hand, if the IPS is based on inertial sensors, the position of the mobile unit is estimated iteratively, based on one of the dead reckoning methods [12]. Finally, if the IPS uses both wireless transceivers and inertial sensors, the position of the mobile node can be computed based on techniques described in the literature as data fusion. Kalman and particle filters are examples of data fusion methods for the estimation of the position of the mobile node. The use of data fusion methods allows increasing the overall accuracy of the IPS, since the combination of different technologies allows minimizing the impact of the drawbacks of each technology. The referred localization methods will be further analyzed in Section IV-B. 可以采用基于统计模型的方法来过滤噪声，从而提高系统的精确度 [35]，[48]。在实际室内环境中，信号会受到无数干扰源的影响，从而产生测量噪声。另一方面，如果 IPS 以惯性传感器为基础，则移动装置的位置将根据其中一种惯性推算法进行迭代估计 [12]。最后，如果 IPS 同时使用无线收发器和惯性传感器，移动节点的位置可根据文献中描述的数据融合技术进行计算。卡尔曼滤波器和粒子滤波器就是估算移动节点位置的数据融合方法。使用数据融合方法可以提高 IPS 的总体精度，因为不同技术的结合可以最大限度地减少每种技术缺点的影响。第四节 B 部分将进一步分析所提及的定位方法。
According to the survey conducted by Li et al. [51], the accuracy of the location information is considered as the most important requirement among first responders in the United States. However, the accuracy of the IPS depends on the information provided. 根据 Li 等人的调查[51]，位置信息的准确性被认为是美国急救人员最重要的要求。然而，IPS 的准确性取决于所提供的信息。
If the IPS provides a room-level position information, its accuracy will be measured by the ratio of correct roomlevel estimations to the total number of targets [52]. For emergency responders the room-level accuracy must be near 100%100 \% [12], [22], [51], otherwise, it will lead to distrust on the IPS. The wrong identification of a room could lead to missing the search in one room or search the same room twice, making the mission inefficient. Furthermore, if an emergency responder is in distress, the wrong identification of his/her room will send the rescue team to a different room, increasing the time of the rescue mission. 如果 IPS 提供的是房间级别的位置信息，其准确性将通过正确的房间级别估计与目标总数的比率来衡量 [52]。对于应急响应人员来说，房间级精度必须接近 100%100 \% [12], [22], [51]，否则会导致对 IPS 的不信任。对房间的错误识别可能会导致错过对一个房间的搜索或对同一房间搜索两次，从而使任务效率低下。此外，如果紧急救援人员遇险，对其房间的错误识别会将救援小组送往不同的房间，增加救援任务的时间。
If the IPS provides the position of the emergency responders as Cartesian coordinates its accuracy will be measured by the average error distance of meter-level estimation of all targets [51]. For such systems, the error in the horizontal and vertical planes should not be, preferably, higher than one and two meters, respectively [17], [28]. Larger errors may be acceptable if they occur occasionally and the system is able to robustly predict them [19]. The need for a robust prediction of the position error leads to another key requirement in the design of an IPS for emergency responders, the reliability of the position estimation. The IPS must be able to recognize when the position estimate is not accurate and notify the emergency responder or the incident commander about it [17], [19], [28]. 如果 IPS 以笛卡尔坐标的形式提供应急响应人员的位置，其精度将通过所有目标的米级估计平均误差距离来衡量 [51]。对于此类系统，水平面和垂直面的误差最好分别不超过 1 米和 2 米 [17]，[28]。如果偶尔出现较大误差，而系统又能对其进行稳健预测，则可以接受[19]。对位置误差进行稳健预测的需求导致了设计用于应急响应者的 IPS 的另一个关键要求，即位置估计的可靠性。IPS 必须能够识别位置估计不准确的情况，并通知应急响应人员或事故指挥官 [17]、[19]、[28]。
B. Information Accessibility B.信息无障碍
Knowing the exact position of the emergency responders is useless if this information in not clearly presented to both the incident commander and the emergency responders. So, the positioning data should be presented intuitively and must not interfere with on-duty activities [19], [28], [51]. Preferably, the positioning information should be presented in smart glasses or helmet visor, in the case of emergency responders. For the incident commander, the positioning information of the emergency responders should be presented in a user-friendly graphical user interface (GUI) [17]. 如果不向事件指挥员和应急响应人员清楚地介绍这些信息，那么了解应急响应人员的确切位置就毫无用处。因此，定位数据应该直观地呈现，并且不得干扰值班活动 [19]、[28]、[51]。对于应急响应人员，定位信息最好通过智能眼镜或头盔面罩呈现。对于事故指挥员来说，应急响应人员的定位信息应在用户友好的图形用户界面（GUI）上显示[17]。
Moreover, the position of the emergency responders must be computed in real-time and the positioning information must be constantly accessible for both, the incident commander and the emergency responders [28]. Each IPS sets a different position update rate. For example, the work carried out by Harmer et al. [17] sets the position update rate to one second and Femminella and Reali [18] specify the minimum position refresh time to 20 seconds. However, according to the survey conducted by Li et al. [52] among the first responders in the 此外，应急响应人员的位置必须实时计算，事件指挥员和应急响应人员必须能够持续获取定位信息[28]。每个 IPS 都设定了不同的位置更新率。例如，Harmer 等人的研究[17] 将位置更新率设定为一秒，Femminella 和 Reali [18] 则将最小位置刷新时间设定为 20 秒。然而，根据 Li 等人[52]对《世界卫生报告》中的第一反应者所做的调查，位置更新率为 1 秒。
United States, the position estimation update should not take more than 40.35 seconds. 在美国，位置估计更新时间不应超过 40.35 秒。
To increase the situational awareness of the incident commander about the mission, the IPS should also have built-in real-time mapping capabilities [19], [28]. The generation of these maps should be automatic and based on Simultaneous Localization and Mapping (SLAM) approaches and emergency responders’ position information. Besides the improvement in situational awareness, these maps can also be used to increase the accuracy of the IPS [19]. 为提高事件指挥官对任务的态势感知能力，IPS 还应具备内置的实时地图绘制功能[19]、[28]。这些地图应根据同步定位和绘图（SLAM）方法以及应急响应人员的位置信息自动生成。除了提高态势感知能力外，这些地图还可用于提高 IPS 的精确度 [19]。
Finally, the information security is another concern in information accessibility. The use of the emergency responders’ position by unauthorized personnel can put them in lifethreatening situations. The access of position information by unauthorized personnel is extremely critical in the military domain. E.g., if the enemies hack the IPS and get access to the positioning information of the militaries on the field, they will be in an advantageous position. Methods for encrypting the voice communications and data transfer can be used to prevent the hacking of the IPS [28]. 最后，信息安全是信息无障碍的另一个关注点。未经授权的人员使用应急响应人员的位置信息可能会危及他们的生命。在军事领域，未经授权人员对位置信息的访问极为重要。例如，如果敌人入侵 IPS 并获取战场上军队的定位信息，他们将处于有利地位。可以使用加密语音通信和数据传输的方法来防止 IPS 遭到黑客攻击[28]。
C. System Adaptability C.系统适应性
As stated before, the missions of emergency responders take place in a vast range of different scenarios. So, to have a good acceptance among the emergency responders, the performance of the IPS should be the same for all intervention scenarios. 如前所述，应急响应人员的任务是在各种不同的场景下执行的。因此，要想获得应急响应人员的认可，IPS 在所有干预场景下的性能都应相同。
Within the research community, it is commonly accepted that an IPS for emergency responders must be (1) independent of building infrastructure (cabling, electricity power supply, and permanent devices installed in fixed locations) [12], [18]-[20], [28], [52], [53], (2) independent of prior collected data [12], [18], [19], [52], [53], (3) independent of accurate maps [19], and (4) robust against dynamic changes on the environment, i.e., without compromising the IPS performance [12], [18]-[20], [22]. These are prerequisites for an IPS to operate in unstructured environments. 研究界普遍认为，用于应急响应人员的 IPS 必须 (1) 不受建筑基础设施（布线、电力供应和安装在固定位置的永久设备）的影响 [12]、[18]-[20]、[28]、[52]、[53]，(2) 不受先前收集的数据的影响 [12]、[18]、[19]、[52]、[53]，(3) 不受精确地图的影响 [19]，(4) 对环境的动态变化具有鲁棒性，即、不影响 IPS 性能 [12]、[18]-[20]、[22]。这些都是 IPS 在非结构化环境中工作的先决条件。
Another important issue about the adaptability of the IPS is the capability to expand its coverage over the site [17]. Expandable coverage is an important feature of an IPS for emergency responders, as their missions can cover wide operative areas. Therefore, if it does not have the capability of expanding its coverage as users enter a building, the IPS will not be able to produce new position updates. 关于 IPS 适应性的另一个重要问题是扩大其覆盖范围的能力[17]。对于应急响应人员来说，可扩展的覆盖范围是 IPS 的一个重要特征，因为他们的任务可能覆盖广泛的工作区域。因此，如果 IPS 不具备在用户进入建筑物时扩大覆盖范围的能力，就无法生成新的位置更新。
However, the deployment of an IPS should not interfere with on-duty activities of emergency responders. Again, according to the survey conducted by Li et al. [52] among the first responders in the United States, the deployment effort on the scene should not take more than 135 seconds. This time represents the period between the moment the emergency responders arrive at the site and the instant the IPS is fully operational. 不过，部署 IPS 不应影响应急人员的执勤活动。同样，根据 Li 等人[52] 对美国急救人员进行的调查，现场部署工作不应超过 135 秒。这一时间是指从应急人员到达现场到 IPS 完全投入运行的时间。
D. System Architecture D.系统架构
One of the main requirements of the IPS architecture is modularity [28]. A modular architecture has many advantages, such as: (1) facilitates the integration with existing systems, (2) reduces modules’ size, therefore, easing the integration on IPS 架构的主要要求之一是模块化[28]。模块化架构有许多优点，例如(1) 便于与现有系统集成，(2) 减少模块的大小，因此便于集成到现有系统中，(3) 减少了系统的成本，(4) 减少了系统的维护成本。
the protective garment and (3) allows the customization of the IPS based on the scenario needs, e.g., if the firefighters are called for a forest fire, the use of a good GPS receiver is enough to meet the localization requirements. This leads to a reduction in energy consumption, expanding the system’s autonomy and making it lighter. 例如，如果消防员被叫去扑灭森林大火，使用一个好的 GPS 接收器就足以满足定位要求。这样就可以减少能源消耗，扩大系统的自主性，使其更加轻便。
Alongside with the capability of expanding the coverage area, an IPS must also be scalable at the number of emergency responders it can support. Scalability is extremely important because the IPS must track all positions of the emergency responders on the field [12], [18], [52]. The number of users to be tracked, range from 100 up to 1000 , depending on the requirements that each developer sets for its IPS [17], [18]. Additionally, the majority of nodes to be tracked are mobile and their number is small when compared with WSNs [12], [22]. The mobility and the reduced number of sensor nodes have implications on the selection of the localization method. On the one hand, the mobility implies that the position of the node has to be continuously computed, forcing the localization algorithm to be computationally tractable, realtime executable and stable to the dynamic changes [12], [20]. On the other hand, the low number of nodes over a wide operative area implies that they have low connectivity, making the use of localization methods of WSNs unreliable for emergency responder scenarios [12], [22]. 除了扩大覆盖范围的能力外，IPS 还必须具有可扩展性，以支持应急响应人员的数量。可扩展性极为重要，因为 IPS 必须跟踪现场应急响应人员的所有位置 [12]、[18]、[52]。需要跟踪的用户数量从 100 到 1000 不等，这取决于每个开发人员对其 IPS 的要求[17], [18]。此外，与 WSN 相比，需要跟踪的大多数节点都是移动节点，而且数量较少 [12], [22]。传感器节点的移动性和数量减少对定位方法的选择有影响。一方面，移动性意味着必须持续计算节点的位置，这就迫使定位算法必须具有可计算性、实时执行性和对动态变化的稳定性 [12]，[20]。另一方面，在广阔的工作区域内，节点数量少意味着它们的连通性低，使得 WSN 的定位方法在应急响应场景中的使用不可靠[12]，[22]。
Since an IPS for emergency responders has to operate under harsh conditions, including extreme temperature and humidity, the physical robustness of its components is a key factor [28], [52]. So, the components of the IPS have to be designed to resist the intense heat, water and other physical hazards, without letting the variation of these environmental parameters affect the performance of the IPS. However, for the ease of assembling on the emergency responder’s garment and to not interfere with the emergency responder’s on-duty activities, these components have to be lightweight and with a reduced form-factor [19], [22], [28], [52]. Based on the survey conducted in the United States among first responders by Li et al. [52], the maximum weight and volume of the IPS are 1.16 Kg and 107.34cm^(3)107.34 \mathrm{~cm}^{3}, respectively. These findings were corroborated by Rantakokko et al. [28] work. 由于用于应急响应的 IPS 必须在极端温度和湿度等恶劣条件下运行，其组件的物理坚固性是一个关键因素[28]，[52]。因此，IPS 组件的设计必须能够抵御高热、水和其他物理危害，而不会让这些环境参数的变化影响 IPS 的性能。但是，为了便于在应急响应人员的服装上装配，并且不影响应急响应人员的执勤活动，这些组件必须重量轻、外形尺寸小[19]、[22]、[28]、[52]。根据 Li 等人在美国对急救人员进行的调查[52]，IPS 的最大重量和体积分别为 1.16 千克和 107.34cm^(3)107.34 \mathrm{~cm}^{3} 。Rantakokko 等人[28] 的研究也证实了这些结论。
E. System Autonomy E.系统自主性
Since the IPS must not be powered by the grid, it has to be battery-powered to provide the IPS the required autonomy. However, depending on the type of emergency, the intervention time can last for long periods, up to 24 hours [28]. Additionally, depending on the localization methods selected, considerable computational resources may be required, which means an increase in power consumption. This entails a tradeoff between IPS’s autonomy and size of the battery or batteries, as an increase of the battery capacity corresponds to an increase in size and weight, which is not desirable. 由于 IPS 不能由电网供电，因此必须由电池供电，以便为 IPS 提供所需的自主性。然而，根据紧急情况的类型，干预时间可能会持续很长时间，最长可达 24 小时[28]。此外，根据所选的定位方法，可能需要大量的计算资源，这意味着耗电量的增加。这就需要在 IPS 的自主性和电池体积之间做出权衡，因为电池容量的增加会导致体积和重量的增加，这是不可取的。
So, energy-efficient methods have to be adopted to cope with the limited energy capacity of the batteries and the computational resources required by the localization methods. These methods focus on the efficient distribution of the computation workload between the emergency responder and the 因此，必须采用节能方法来应对电池有限的能量容量和定位方法所需的计算资源。这些方法的重点是在应急响应者和定位系统之间有效分配计算工作量。
incident commander, and to keep the communication overhead low [12]. In this way, at the emergency responder side, the first estimation of emergency responder position is carried out with the help of computationally tractable algorithms. This information is then sent to the incident commander, where powerful algorithms will refine the position of the emergency responder based on the position and the spatial distribution of the other emergency responders and building maps. The used building maps can be based on constructed virtual maps or on the existing building maps. 这样，在应急响应方，可借助可计算的算法对应急响应方的位置进行首次估计。这样，在应急响应方一侧，借助可计算的算法对应急响应方的位置进行首次估计。然后，这些信息会被发送给事故指挥官，在那里，功能强大的算法会根据其他应急响应者的位置和空间分布情况以及建筑地图来完善应急响应者的位置。使用的建筑地图可以是基于构建的虚拟地图，也可以是基于现有的建筑地图。
F. Cost F.费用
The cost of the IPS has a huge impact on its acceptance by the emergency responders [12], [19], [28], [52]. Due to the challenging scenarios encountered during these missions, some researchers developed prototype and innovative technologies to address this issue. However, due to the early stage of their development, these technologies are quite expensive, which result in an expensive IPS. According to Rantakokko et al. [28] to achieve a high market penetration, the price of an IPS should be below undefined. Nevertheless, this value may vary between countries. IPS 的成本对应急响应人员的接受程度有很大影响 [12]、[19]、[28]、[52]。由于在执行这些任务时会遇到具有挑战性的情况，一些研究人员开发了原型和创新技术来解决这一问题。然而，由于处于早期开发阶段，这些技术相当昂贵，导致 IPS 造价高昂。根据 Rantakokko 等人的观点[28]，要实现较高的市场渗透率，IPS 的价格应低于 undefined 。尽管如此，这一数值可能因国家而异。
IV. Localization Techniques and Methods IV.定位技术和方法
Localization is the ability to know, at a given time, the position of a person or object in relation to a predefined referential. The localization problem can be divided into outdoor and indoor localization. This division is justified by some differences at conception level, e.g., the need for a global or local reference, the accuracy required, and the influence of the environment on the precision and accuracy of localization, based on the selected technology. 定位是指在给定时间内，了解人或物体相对于预定参照物的位置的能力。定位问题可分为室外定位和室内定位。这种划分是有道理的，因为在概念层面上存在一些差异，例如，根据所选择的技术，对全局或局部参照物的需求、所要求的精度以及环境对定位精度和准确性的影响。
In the literature, many terms have been used to define the location process of a person/object. If the localization system relies on wireless technologies, terms like localization, positioning, geolocation, position location, and radiolocation are commonly used to describe the location process [31]. 文献中使用了许多术语来定义人/物的定位过程。如果定位系统依赖于无线技术，则通常使用定位、定位、地理定位、位置定位和无线电定位等术语来描述定位过程[31]。
In the following subsections, the distance measurement techniques and the localization methods used in indoor localization are explored. When compared to the works presented in [1], [27], [31], [36], [38], and [43], a wider selection of localization techniques and methods are addressed and analyzed, focusing on their theoretical background and the pros and cons of their use. The analysis presented provides a good theoretical background on the indoor localization field, as well as, a better understanding of the main limitations affecting the performance of the developed IPSs for emergency responders. The schemes proposed for emergency responders’ IPSs, discussed in Section VII, are based on one or the combination of two or more techniques and methods described below. 以下各小节将探讨室内定位中使用的距离测量技术和定位方法。与文献[1]、[27]、[31]、[36]、[38]和[43]中介绍的工作相比，本节探讨和分析了更多的定位技术和方法，重点是它们的理论背景以及使用这些技术和方法的利弊。所做的分析为室内定位领域提供了良好的理论背景，并使人们更好地了解影响应急响应者所开发的 IPS 性能的主要限制因素。第七节讨论的为应急响应人员 IPS 提出的方案是基于下述一种或两种以上技术和方法的组合。
The localization methods of WSNs are not covered as they do not comply with the IPS requirements for emergency responders. WSN 的定位方法不包括在内，因为它们不符合 IPS 对应急响应人员的要求。
A. Distance Measurement Techniques A.距离测量技术
The selection of the distance measurement technique or combination of techniques is a key factor for the performance 距离测量技术或技术组合的选择是影响性能的关键因素。
Fig. 3. Taxonomy of the Distance Measurement Techniques. 图 3.距离测量技术分类法。
of the IPS. Thereby, the accuracy of an IPS will be directly related to the accuracy of the distance measured. The same applies for the system’s precision. 因此，IPS 的精度与测量距离的精度直接相关。因此，IPS 的精度与测量距离的精度直接相关。系统的精度也是如此。
Several technologies have been used for distance measurement between nodes, e.g., ultrasound sensors [22], [33], [35], [39], [41], [43]; infrared sensors [33], [35], [39], [43]; video cameras [33], [37], [39]; inertial or motion sensors [12], [22], [40], [42]; techniques based on radio frequency signals [12], [22], [31], [33], [35], [39], [43]; or a combination of these technologies [12], [22], [26], [39], [43]. However, some of these systems showed performance limitations due to the restriction of line-of-sight (LOS). To overcome the LOS restriction, some IPSs require beforehand knowledge of the sensor orientation or rely on targeting systems, which are usually very expensive and require enormous processing time. Therefore, in systems that the human interaction should be nonexistent and the power consumption is a critical factor, the use of directional sensors must be avoided. In fact, designing an omnidirectional and accurate system is a highly complex problem that continues to be a major topic of research worldwide [54]. 节点之间的距离测量采用了多种技术，例如超声波传感器 [22]、[33]、[35]、[39]、[41]、[43]；红外传感器 [33]、[35]、[39]、[43]；摄像机 [33]、[37]、[39]；惯性或运动传感器 [12]、[22]、[40]、[42]；基于射频信号的技术 [12]、[22]、[31]、[33]、[35]、[39]、[43]；或这些技术的组合 [12]、[22]、[26]、[39]、[43]。然而，由于视线（LOS）的限制，其中一些系统的性能受到了限制。为了克服视距限制，一些 IPS 需要事先了解传感器的方位，或者依赖于瞄准系统，而瞄准系统通常非常昂贵，需要大量的处理时间。因此，在不存在人机交互且功耗是关键因素的系统中，必须避免使用定向传感器。事实上，设计一个全向和精确的系统是一个非常复杂的问题，目前仍是全球研究的一个主要课题[54]。
According to the distance measurement principle, these techniques can be classified into two categories based on the (1) properties of wireless communication technologies and on (2) measurement of physical quantities. The taxonomy of the distance measurement techniques is shown in Fig. 3. 根据距离测量原理，这些技术可按照（1）无线通信技术的特性和（2）物理量的测量分为两类。距离测量技术的分类如图 3 所示。
Techniques Based on the Properties of Wireless Communication Technologies: The techniques based on the wireless communication properties can be divided into four categories: (1) radio hop count; (2) signal propagation time (ToA, TDoA, and RToF); (3) received signal phase/angle (AoA, RSP, and interferometry); and (4) received signal strength (RSS and CSI). 基于无线通信技术特性的技术：基于无线通信特性的技术可分为四类：(1) 无线电跳数；(2) 信号传播时间（ToA、TDoA 和 RToF）；(3) 接收信号相位/角度（AoA、RSP 和干涉测量）；以及 (4) 接收信号强度（RSS 和 CSI）。
The Radio Hop Count technique assumes that the network nodes are evenly distributed and the distance between the nodes is known beforehand. Thus, the distance between two nodes is defined by the number of hops needed to establish the communication between them. However, to achieve a good performance, this technique requires a high 无线电跳数技术假定网络节点均匀分布，节点之间的距离事先已知。因此，两个节点之间的距离由它们之间建立通信所需的跳数定义。不过，要实现良好的性能，这种技术需要较高的
node density, which is unrealistic in emergency responder’s scenarios [12], [22], [38]. 节点密度，这在应急响应场景中是不现实的 [12]、[22]、[38]。
The Time of Arrival (ToA), also known as Time of Flight ( ToF ), is based on the fundamental principle that the distance between the transmitter and the receiver is directly proportional to the transmission time delay and the signal velocity [35]. In the air the speed of electromagnetic waves is approximately 3.10^(8)m//s3.10^{8} \mathrm{~m} / \mathrm{s}. The major drawback of this technique is the need of a high precision temporal synchronization between the transmitter and the receiver. A synchronization error of one nanosecond results in an error of 0.3 meters in the localization [55]. Nevertheless, the ToA technique is very accurate and can filter out the multipath effects [33]. The best results are achieved when the ToA technique is combined with UWB technology [35]. 到达时间（ToA）又称飞行时间（ToF），其基本原理是发射器和接收器之间的距离与传输时延和信号速度成正比 [35]。在空气中，电磁波的速度约为 3.10^(8)m//s3.10^{8} \mathrm{~m} / \mathrm{s} 。这种技术的主要缺点是发射器和接收器之间需要高精度的时间同步。一纳秒的同步误差会导致 0.3 米的定位误差 [55]。不过，ToA 技术非常精确，可以滤除多径效应 [33]。将 ToA 技术与 UWB 技术相结合可获得最佳效果 [35]。
The Time Difference of Arrival (TDoA) technique, also known as Time Difference of Flight (TDoF), is a variation of the ToA technique and has two approaches: the TDoA of multiple nodes [12] and the TDoA of multiple signals [35], [38]. In the first approach, several anchor nodes (receivers) are distributed on the site and temporally synchronized. Periodically, the target node (transmitter) broadcasts a signal. When this signal is received by an anchor node, the timestamp of the signal reception is recorded. The recorded values are sent to a base station that is responsible for the computation of the target node position. This technique can be applied in reverse. In this case, the anchor nodes transmit, simultaneously, the ranging signal. The target node receives those signals and computes its position. In the second approach, two signals are transmitted simultaneously but with different propagation velocities in the medium. As the propagation speed of those signals is known, the distance between two devices is computed based on the time difference of arrival of the two signals at the receiver [35], [38]. Like the ToA technique, the TDoA technique is only suitable when LOS is available. The position of the anchor nodes has to be known beforehand, and they have to share the same temporal reference [12], [31], [43]. The advantage of this technique over the ToA technique is that the target node does not need to be synchronized with the anchor nodes [31], [43]. The TDoA variant with multiple signals usually requires extra hardware [38]. 到达时间差（TDoA）技术，又称飞行时间差（TDoF），是 ToA 技术的一种变体，有两种方法：多节点 TDoA [12] 和多信号 TDoA [35]、[38]。在第一种方法中，多个锚节点（接收器）分布在站点上，并在时间上同步。目标节点（发射器）定期广播信号。当一个锚节点接收到该信号时，信号接收的时间戳就会被记录下来。记录的值被发送到负责计算目标节点位置的基站。这种技术也可反向应用。在这种情况下，锚节点同时发射测距信号。目标节点接收这些信号并计算其位置。在第二种方法中，两个信号同时传输，但在介质中的传播速度不同。由于这些信号的传播速度是已知的，因此可以根据两个信号到达接收器的时间差来计算两个设备之间的距离 [35]，[38]。与 ToA 技术一样，TDoA 技术只适用于有 LOS 的情况。必须事先知道锚节点的位置，而且它们必须共享相同的时间参考 [12]、[31]、[43]。与 ToA 技术相比，该技术的优势在于目标节点无需与锚节点同步 [31], [43]。带有多个信号的 TDoA 变体通常需要额外的硬件 [38]。
The Round-Trip Time of Flight (RToF) technique, which is also based on the ToA technique, consists in measuring the time that a signal takes to travel the path between a transmitter and a receiver and return to the transmitter. The processing time that a receiver requires to process the incoming signal and send it back is subtracted to the time measured by the transmitter [12], [35]. Unlike the ToA and the TDoA techniques, this technique does not require the clock synchronization among the nodes [12], [31]. The localization methods used in ToA-based systems can also be applied to this technique, without adaptation [31]. However, the RToF technique requires high precision clocks for the computation of the receiver processing time, which increase the cost of the system [12], [31]. The communication overhead and the complexity of the system will also increase due the messages exchange [12], [35], [43]. 往返飞行时间（RToF）技术也是以 ToA 技术为基础的，它包括测量信号在发射器和接收器之间的路径上传输并返回发射器所需的时间。接收器处理输入信号并将其发送回来所需的处理时间减去发射器测量的时间[12], [35]。与 ToA 和 TDoA 技术不同，这种技术不需要节点之间的时钟同步 [12], [31]。基于 ToA 的系统中使用的定位方法也可应用于这种技术，无需进行调整 [31]。不过，RToF 技术需要高精度时钟来计算接收器处理时间，这增加了系统成本 [12]，[31]。由于信息交换，通信开销和系统复杂性也会增加 [12]、[35]、[43]。
The Angle of Arrival (AoA) technique consists of determining the reception angle of a received signal. In AoA-base systems, the position of the anchor nodes must be previously known [31], [43], and the computation of the received angle can either be performed by the target node or the anchor node [12]. For the determination of the angle of arrival, the system can either use directional antennas or antennas arrays [31], [35], [38], [43]. The AoA technique does not need temporal synchronization among nodes [31], [35]. However, to achieve a good localization accuracy, the AoA technique needs LOS availability [12], [31], [35], [55]. Additionally, the IPS has to be based on antennas with angles’ measurement capability, which will increase the cost of the system [35], [43], [55]. 到达角（AoA）技术包括确定接收信号的接收角。在基于 AoA 的系统中，必须事先知道锚节点的位置 [31]、[43]，接收角的计算可以由目标节点或锚节点完成 [12]。为了确定到达角，系统可以使用定向天线或天线阵列 [31], [35], [38], [43]。AoA 技术不需要节点间的时间同步 [31], [35]。然而，要实现良好的定位精度，AoA 技术需要有可用的 LOS [12]、[31]、[35]、[55]。此外，IPS 必须基于具有角度测量能力的天线，这将增加系统成本 [35]、[43]、[55]。
The Received Signal Phase (RSP) technique uses the carrier phase for distance calculation. This technique presupposes the transmission of signals at the same frequency, purely sinusoidal and with zero offsets of phase. For the determination of the arrival phase, the signals must be transmitted with a finite propagation delay and the transmitters’ position must be known [31]. However, this technique needs LOS conditions to work [31]. 接收信号相位（RSP）技术使用载波相位来计算距离。该技术的前提条件是传输相同频率、纯正弦波且相位偏移为零的信号。为了确定到达相位，信号的传输必须有有限的传播延迟，并且必须知道发射器的位置[31]。不过，这种技术需要在 LOS 条件下才能发挥作用 [31]。
The Interferometry technique consists of transmitting sinusoidal waves with multiple frequencies. The receiver, equipped with antennas arrays, stores the different overlapped signals. Based on the phase displacements and knowing the signals wavelength, the distance between two devices can be computed [12]. This technique also requires LOS to compute the distance accurately [12]. The need of LOS conditions is one of the main reasons to not recommend the use of techniques based on received signal phase/angle for the indoor localization [55]. 干涉测量技术包括发射多个频率的正弦波。配备天线阵列的接收器存储不同的重叠信号。根据相位差和已知的信号波长，可以计算出两个设备之间的距离 [12]。这种技术也需要 LOS 才能准确计算距离 [12]。需要 LOS 条件是不建议使用基于接收信号相位/角度的室内定位技术的主要原因之一 [55]。
The Received Signal Strength (RSS) technique is based on the physical principle that the received strength of a radio signal is inversely proportional to the distance increase. To compute the distance between nodes, the IPS has to have a prior knowledge about the signal transmission power [12]. However, in indoor environments, the relationship between the RSS and the transmission power may be nonlinear. This nonlinearity occurs due to the countless obstructions that a signal is subjected to in indoor environments, as well as, to the effects of phenomena like multipath, scattering, diffraction and reflection in a radio signal [55]. In the works conducted by Sarkar et al. [56] and Zanella [57], several empirical and theoretical models are analyzed aiming to minimize the error sources associated with the aforementioned problems. The RSS technique allows reusing the existing infrastructure since most of the wireless technologies have the built-in capability of RSS measurement [12], [43], [55]. Additionally, this technique works under non-line-of-sight (NLOS) conditions when a propagation model or a fingerprinting-based localization method is used. However, the signal propagation models and the fingerprinting-based methods are site specific. The main drawback of the RSS technique is the fluctuation of the RSS value between readings [12], [31], [43], [55], [58]. 接收信号强度（RSS）技术的物理原理是，无线电信号的接收强度与距离的增加成反比。为了计算节点之间的距离，IPS 必须事先了解信号传输功率[12]。然而，在室内环境中，RSS 与传输功率之间的关系可能是非线性的。出现这种非线性的原因是信号在室内环境中受到无数障碍物的影响，以及无线电信号中的多径、散射、衍射和反射等现象的影响 [55]。在 Sarkar 等人[56] 和 Zanella [57]的研究中，分析了多个经验和理论模型，旨在最大限度地减少与上述问题相关的误差源。RSS 技术允许重复使用现有的基础设施，因为大多数无线技术都具有内置的 RSS 测量功能 [12]、[43]、[55]。此外，当使用传播模型或基于指纹的定位方法时，该技术可在非视距（NLOS）条件下工作。不过，信号传播模型和基于指纹的方法都是针对具体地点的。RSS 技术的主要缺点是读数之间 RSS 值的波动 [12]、[31]、[43]、[55]、[58]。
The Channel State Information (CSI) technique relies on the PHY layer power feature to discriminate multipath characteristics of a radio signal [59]. This technique studies the channel variation experienced during the propagation of a signal, more 信道状态信息（CSI）技术依靠物理层功率特征来区分无线电信号的多径特性[59]。该技术研究信号传播过程中经历的信道变化，更多的是研究信号的多径特性。
specifically, reveals the influence of scattering, fading and power decay with distance [60], [61]. Unlike RSSI, CSI is able to capture and characterize all the individual components of the multipath effect [59]. To study the multipath components, two variants of the CSI technique can be used, namely, the Channel Impulse Response (CIR) and the Channel Frequency Response (CFR). CIR and CFR are, respectively, the time and frequency domain representation of the individual multipath components. The CIR variant allows separating the LOS component, while the CFR variant is useful to study the frequency-selective fading [59]. When compared with RSSI-based approaches, the main advantages of CSI technique are the capability to distinguish the multipath components and the possibility of using the same localization methods of RSS-based systems, with higher accuracy in both CSI-assisted ranging and CSI-assisted fingerprinting [59]. However, until now there is only one commercial off-theshelf (COTS) device that allows extracting the CSI values from the PHY layer. This device only works with Orthogonal Frequency Division Multiplexing (OFDM) based system and requires a site survey phase when it is used with CSI-assisted fingerprinting [59], [61]. 具体而言，它揭示了散射、衰减和功率随距离衰减的影响 [60]，[61]。与 RSSI 不同，CSI 能够捕捉并描述多径效应的所有单独成分 [59]。为了研究多径成分，可以使用 CSI 技术的两种变体，即信道脉冲响应（CIR）和信道频率响应（CFR）。CIR 和 CFR 分别是各个多径成分的时域和频域表示。CIR 变体可以分离 LOS 分量，而 CFR 变体则有助于研究频率选择性衰落 [59]。与基于 RSSI 的方法相比，CSI 技术的主要优势在于能够区分多径分量，并且可以使用与基于 RSS 的系统相同的定位方法，在 CSI 辅助测距和 CSI 辅助指纹识别中都具有更高的精度[59]。不过，到目前为止，只有一种商用现成（COTS）设备可以从 PHY 层提取 CSI 值。这种设备仅适用于基于正交频分复用（OFDM）的系统，在与 CSI 辅助指纹识别一起使用时，需要进行现场勘测[59]，[61]。
2) Techniques Based on the Measurement of Physical Quantities: The distance determination through the measurement of physical quantities can be carried out by two different techniques: direct measurement and use of inertial and motion sensors. 2) 基于物理量测量的技术：通过测量物理量来确定距离可采用两种不同的技术：直接测量和使用惯性传感器和运动传感器。
The Direct Measurement technique is based on using physical actions (e.g., contact or pressure) or movements (with the arm of a robot, for instance) to compute the distance between the user and the device. The main drawbacks of this technique are the requirement for LOS and the need for interaction between the user and the system. 直接测量技术的基础是使用物理动作（如接触或压力）或运动（如使用机器人手臂）来计算用户与设备之间的距离。这种技术的主要缺点是需要 LOS 以及用户与系统之间的互动。
The use of Inertial and Motion Sensors is another technique for the measurement of physical quantities. This class of proprioceptive sensors is capable of measuring physical phenomena such as acceleration, velocity and orientation of the movement [12] and allows the computation of the traveled distance and the orientation of the movement. The direct approach computes the displacement and the heading changes between steps through the direct use of magnetometers, accelerometers and gyroscopes readings. In the indirect approach, the readings are used to detect and count the steps, as well as, the heading changes. Based on the number of steps and the distance of each step, the movement displacement is computed. The distance of each step can be a fixed value or a value computed iteratively from the inertial sensors reading [40], [42]. This technique is fully autonomous and, therefore, no external device is required for the computation of the traveled distance. This is the major advantage of this technique over the other distance measurement techniques. However, the main drawback of this technique is the error control associated with each measurement, since it presents a cumulative behavior [12], [43]. 惯性和运动传感器是测量物理量的另一种技术。这类本体感觉传感器能够测量加速度、速度和运动方向等物理现象[12]，并能计算移动距离和运动方向。直接方法是通过直接使用磁力计、加速度计和陀螺仪读数来计算两步之间的位移和方向变化。在间接方法中，读数用于检测和计算步数以及方向变化。根据步数和每一步的距离，计算出运动位移。每一步的距离可以是一个固定值，也可以是根据惯性传感器读数迭代计算得出的值 [40], [42]。这种技术是完全自主的，因此不需要外部设备来计算移动距离。与其他测距技术相比，这是该技术的主要优势。不过，这种技术的主要缺点是每次测量的误差控制，因为它呈现累积行为 [12], [43]。
B. Localization Methods B.定位方法
In an IPS, the localization method is the core of the system and the unit responsible for computing the position 在 IPS 中，定位方法是系统的核心，也是负责计算位置的单元。
Fig. 4. Taxonomy of Localization Methods. 图 4.定位方法分类。
of a target node. The position is computed based on the distance measurements carried out by one or more of the techniques described in the previous subsection. Besides computing the target node’s position, the localization methods can also rely on measurement error mitigation mechanisms to improve the localization accuracy. These mitigation mechanisms are designed to deal with the physical phenomena that affect radio signals - e.g., multipath, reflection, scattering, obstruction, and attenuation - or the drift error of the inertial sensors. A common practice to increase the IPS performance is the combination of two or more localization methods. 目标节点的位置。位置是根据上一小节所述的一种或多种技术进行的距离测量计算得出的。除了计算目标节点的位置，定位方法还可以依靠测量误差缓解机制来提高定位精度。这些减缓机制旨在处理影响无线电信号的物理现象（如多径、反射、散射、阻挡和衰减）或惯性传感器的漂移误差。提高 IPS 性能的常见做法是将两种或多种定位方法结合起来。
However, considering the requirements for emergency responders’ IPS presented Section III-C, a new taxonomy to classify the localization methods is proposed, Fig. 4. This taxonomy aims to stress the restrictions related to the selection of a localization method and to help new researchers in the field to avoid the pitfalls encountered by previous researchers. In the proposed taxonomy, the localization methods are classified according to the need of a training or offline phase. Based on this classification two classes of localization methods are reviewed, the (1) autonomous and the (2) training-dependent methods. 不过，考虑到第 III-C 节介绍的应急响应人员 IPS 的要求，提出了一种新的定位方法分类法（图 4）。该分类法旨在强调与选择定位方法有关的限制，并帮助该领域的新研究人员避免之前研究人员遇到的陷阱。在建议的分类法中，本地化方法根据是否需要训练或离线阶段进行分类。在此分类法的基础上，评述了两类本地化方法：(1) 自主方法和 (2) 依赖训练的方法。
Autonomous Methods: The autonomous methods can be divided into four categories, namely, (1) proximity methods, (2) geometric approaches (e.g., triangulation, trilateration, and maximum likelihood estimation), (3) dead reckoning methods and (4) Kalman filters. 自主方法：自主方法可分为四类，即：(1) 接近方法；(2) 几何方法（如三角测量法、三坐标法和最大似然估计法）；(3) 惯性推算法；(4) 卡尔曼滤波器。
The Proximity localization method is based on a dense network of antennas distributed throughout the site, whose position is previously known. These antennas can either be active or passive and based on the antenna type, the IPS can be centralized or decentralized, respectively. The position of the mobile device is defined as the detected antenna position. Therefore, the accuracy of the system is inversely proportional to the antennas’ cover range and directly proportional to the granularity of antennas [43]. If a mobile node is detected or detects two antennas, its position is defined by the antenna’s position that has the strongest signal. This method does not 近距离定位方法基于分布在整个站点的密集天线网络，这些天线的位置是已知的。这些天线可以是有源的，也可以是无源的，根据天线的类型，IPS 可以是集中式的，也可以是分散式的。移动设备的位置被定义为检测到的天线位置。因此，系统的精确度与天线的覆盖范围成反比，与天线的粒度成正比[43]。如果一个移动节点被检测到或检测到两根天线，则其位置由信号最强的天线位置定义。这种方法不
allow continuous monitoring of the position. Proximity is widely used in combination with dead reckoning methods to correct the mobile node position. This strategy is very efficient for error reduction/control [12]. The simplicity and ease of integration of this method with other localization methods make the proximity method very attractive for emergency responders’ IPS [12], [31], [43]. 可对位置进行持续监测。近距离定位法与死位推算法相结合，被广泛用于校正移动节点的位置。这种策略在减少/控制误差方面非常有效 [12]。这种方法与其他定位方法的结合既简单又容易，因此对于应急响应人员的 IPS 来说，接近法非常有吸引力 [12]、[31]、[43]。
The Triangulation or Angulation method relies on angle measurements and the geometric properties of triangles to estimate the mobile node’s position. The angle measurements are obtained through the AoA technique [31], [35]. This method allows computing the 2 D and 3 D position with two and three angle measurements, respectively. For the 3 D position, the third angle measurement is the azimuth angle. The position of the mobile node is computed through the intersection of imaginary lines created from these angle measurements [35], [55], [58]. The accuracy of the IPS can be increased if more angle measurements are collected from other anchor nodes [43]. 三角测量法或角度法依靠角度测量和三角形的几何特性来估计移动节点的位置。角度测量通过 AoA 技术获得 [31]，[35]。这种方法可通过两次和三次角度测量分别计算 2 D 和 3 D 位置。对于 3 D 位置，第三个角度测量值是方位角。移动节点的位置是通过这些角度测量值创建的假想线的交点计算出来的 [35]、[55]、[58]。如果能从其他锚节点收集到更多的角度测量值，IPS 的精确度就能得到提高 [43]。
Like the triangulation method, the Trilateration method also computes the node’s position based on the geometric proprieties of triangles. However, instead of using angle measurements, the trilateration method relies on distance measurements [43]. The 2D or 3D position of the target node is computed based on three non-collinear distance measurements or four non-coplanar distance measurements, respectively. The position of the target node is computed by the intersection of imaginary circles or spheres for 2D or 3D localization, respectively [35], [38], [55]. The ray of each imaginary circle or sphere corresponds to the distances obtained through measurement techniques like ToA, TDoA RToF, RSS, interferometry and CSI [31], [43]. Additionally, the angle measurements obtained from AoA techniques can also be used if the angles are converted into distances [35]. In indoor environments, time-based techniques are more precise than RSS-based techniques [55]. 与三角测量法一样，三坐标法也是根据三角形的几何特性计算节点的位置。不过，三坐标法不使用角度测量，而是依靠距离测量[43]。目标节点的二维或三维位置分别根据三次非共线距离测量或四次非共面距离测量计算得出。在二维或三维定位中，目标节点的位置分别通过虚圆或虚球的交点来计算 [35]、[38]、[55]。每个虚圆或虚球的射线与通过 ToA、TDoA RToF、RSS、干涉测量和 CSI 等测量技术获得的距离相对应[31]，[43]。此外，如果将角度转换为距离，也可以使用通过 AoA 技术获得的角度测量值[35]。在室内环境中，基于时间的技术比基于 RSS 的技术更加精确[55]。
The Maximum Likelihood Estimation (MLE) method is a subtype of the trilateration method and is used to address the problem of measurement uncertainty in localization [35]. The main goal of the MLE method is to minimize the impact that noisy environments have on distance measurements. Based on distance measurements from multiple anchor nodes, usually more than three, a statistical method is used to minimize the differences between the measured distances and estimated distances [35], [38]. Another advantage of the MLE method is the capability to address the temporal synchronization issues [35]. 最大似然估计法（MLE）是三坐标法的一个子类型，用于解决定位中的测量不确定性问题[35]。MLE 方法的主要目标是尽量减少噪声环境对距离测量的影响。根据来自多个锚节点（通常超过三个）的距离测量值，使用一种统计方法来最小化测量距离与估计距离之间的差异 [35]，[38]。MLE 方法的另一个优点是能够解决时间同步问题 [35]。
The Dead Reckoning method consists in determining the current position based on the last computed position and the motion information. The position of the target node is iteratively incremented based on the known/calculated velocity of the motion over a defined period of time [12], [43]. The dead reckoning method is usually applied in systems based on inertial and motion sensors but it can also be used in video-based IPSs. For the estimation of the distance traveled by a person, two different techniques can be used: the step counter technique, called pedometer; and the Pedestrian Dead Reckoning (PDR) technique. The former estimates the 惯性推算法是根据上次计算的位置和运动信息确定当前位置。目标节点的位置会根据已知/计算出的运动速度在规定时间内迭代递增 [12]，[43]。惯性推算法通常应用于基于惯性和运动传感器的系统，但也可用于基于视频的 IPS。在估算人的行走距离时，可以使用两种不同的技术：一种是被称为计步器的计步技术，另一种是行人惯性推算（PDR）技术。前者估算的是
distance traveled by counting the number of steps and then multiplying by the predefined step length. The main drawback of this approach is the choice of an appropriate step length for the computation of the traveled distance. Furthermore, normal walking has non-constant step length which will lead to position estimation errors. The PDR technique estimates the traveled distance directly, i.e., one IMU is attached to the foot of a person and measures the changes in acceleration and angular velocity. The acquired information is then processed, converted into the motion referential and the distance and orientation of the movement are computed. The movement displacement and the orientation change are incremented in the last position computed. To increase the performance of the system, the Zero-Velocity Updates (ZUPT) technique can be used. The ZUPT technique sets the velocity to zero when a stance phase is detected, i.e., when the foot touches the ground. This technique allows reducing the impact of the inertial sensors drifts to one measurement cycle, because it prevents the error associated with the sensor drift from being iteratively integrated. Without the ZUPT technique, the position error will grow exponentially due to the two-fold integration needed to convert the acceleration into a distance [12]. Also, waypoints and maps can be used for error mitigation. Waypoints are based on the proximity localization method and, usually, on the Radio-Frequency Identification (RFID) technology. Another alternative for error mitigation is the use of the context information provided by maps. In this approach, for each new position estimation, the system will check its plausibility according to the existing map floor. The system assumes that a person cannot cross through walls and can only change rooms via doors, which are also detailed on the map. Moreover, since the position of the doors and stairs are known, this information can also be used to correct the position estimation [12]. 计算移动距离的方法是计算步数，然后乘以预先设定的步长。这种方法的主要缺点是计算行走距离时需要选择合适的步长。此外，正常行走的步长并不恒定，这将导致位置估计误差。PDR 技术直接估算行走距离，即在人的脚上安装一个 IMU，测量加速度和角速度的变化。然后对获取的信息进行处理，转换成运动参考值，并计算出运动的距离和方向。运动位移和方向变化在最后计算的位置上递增。为了提高系统的性能，可以使用零速度更新（ZUPT）技术。当检测到站立阶段，即脚接触地面时，ZUPT 技术会将速度设为零。这种技术可以将惯性传感器漂移的影响减少到一个测量周期，因为它可以防止与传感器漂移相关的误差被迭代整合。如果没有 ZUPT 技术，由于将加速度转换为距离需要进行两次积分，位置误差将呈指数增长[12]。此外，航点和地图也可用于减少误差。航点基于近距离定位方法，通常采用射频识别（RFID）技术。另一种减少误差的方法是使用地图提供的上下文信息。在这种方法中，对于每一个新的位置估计，系统都会根据现有的地图底层检查其可信度。 该系统假定人不能穿过墙壁，只能通过门（地图上也有详细说明）来更换房间。此外，由于门和楼梯的位置是已知的，这些信息也可用于修正位置估计 [12]。
The Kalman Filter was first introduced by Rudolph E. Kalman and is a subtype of a Bayesian filter [62]. The Kalman filter is an iterative method for data fusion [48]. This method can be used in any system that relies on distance and bearing measurements between nodes or other indirect measurements, such as velocity, acceleration, and angular velocity [21]. This method is based on the estimation of the state of a linear dynamic system, given a series of observations [63]. A Kalman filter has two steps, the prediction step, and the update step. In the prediction step, the next state of the system is predicted based on the previous measurements. In the update state, the current state of the system is estimated based on the latest measurements acquired [63]. The new position estimation is computed based on the combination of the results of these two steps. For nonlinear dynamic systems, there are two extensions to the Kalman filter for the calculation of the system states, the Extended Kalman Filter (EKF) and the Unscented Kalman Filter (UKF). The difference between these two extensions is the Gaussian approximation to the filtering solutions of nonlinear optimal filtering problems. The EKF uses a Taylor series based approximation while the UKF uses an unscented transform [63]. The Kalman filter has several advantages, namely, good accuracy, robustness even with 卡尔曼滤波器由 Rudolph E. Kalman 首次提出，是贝叶斯滤波器的一个子类型 [62]。卡尔曼滤波器是一种用于数据融合的迭代方法 [48]。这种方法可用于任何依赖节点间距离和方位测量或其他间接测量（如速度、加速度和角速度）的系统 [21]。卡尔曼滤波法基于一系列观测数据对线性动态系统的状态进行估计 [63]。卡尔曼滤波器有两个步骤：预测步骤和更新步骤。在预测步骤中，根据之前的测量结果预测系统的下一个状态。在更新状态下，根据最新获得的测量值估算系统的当前状态 [63]。新的位置估计是根据这两个步骤的结果组合计算得出的。对于非线性动态系统，卡尔曼滤波器有两种用于计算系统状态的扩展方案，即扩展卡尔曼滤波器（EKF）和无标点卡尔曼滤波器（UKF）。这两种扩展的区别在于对非线性优化滤波问题的滤波解进行高斯近似。EKF 使用基于泰勒级数的近似方法，而 UKF 则使用无符号变换 [63]。卡尔曼滤波器有几个优点，即精度高、鲁棒性强，即使在出现以下情况时也不例外
noisy measurements, the capability to track several mobile nodes in real-time and low computational and communication overhead [21]. 噪声测量、实时跟踪多个移动节点的能力以及较低的计算和通信开销 [21]。
2) Training-Dependent Methods: The training-dependent methods can be divided into five categories, namely, (1) centroid, (2) particle filters, (3) pattern recognition, (4) visual analysis and (5) map matching. 2) 依赖训练的方法：依赖训练的方法可分为五类，即 (1) 中心点法、(2) 粒子过滤器法、(3) 模式识别法、(4) 视觉分析法和 (5) 地图匹配法。
The Centroid method is composed of a training phase and an online phase. During the training phase, the position of each anchor node is computed based on the measurements that each anchor node receives from their neighbor nodes. The final position of an anchor node is the result of a weighted average of the position estimated from each measurement. This information is stored as a radio map that contains all anchor nodes of the system. During the online phase, the position of the mobile node is set as being at the center of all hearable anchor nodes. The Cartesian coordinates of the mobile node are computed based on the arithmetic average of the position of each hearable anchor node. Additionally, the position of a mobile node can be further refined using a weighted average approach. This approach is based on the RSS from each anchor node and on attaching a different weight to the position of each anchor node [55]. 中心点法由训练阶段和在线阶段组成。在训练阶段，每个锚节点的位置是根据每个锚节点从其邻居节点接收到的测量结果计算出来的。锚节点的最终位置是根据每次测量估算的位置加权平均的结果。这些信息以无线电地图的形式存储，其中包含系统的所有锚节点。在联机阶段，移动节点的位置被设置为位于所有可听到的锚节点的中心。移动节点的笛卡尔坐标是根据每个可听锚节点位置的算术平均值计算得出的。此外，还可以使用加权平均法进一步完善移动节点的位置。这种方法以每个锚节点的 RSS 为基础，并为每个锚节点的位置附加不同的权重[55]。
The Particle Filter (PF) method is a data fusion algorithm that merges the data provided by different sensors, i.e., combines the data delivered by proprioceptive and exteroceptive sensors. The position of a mobile node, at a given moment, is computed based on a set of weighted particles. A particle is an estimation of the position generated by the PF. An IPS based on this method requires two entry models, the sensorial and the motion model. The sensorial model determines, based on sensorial data, what is the probability of the estimated position (particle) to match with the real position of the device. The motion model describes the particle movement as an approximation to the mobile node movement. The motion can be described either by simple models (e.g., generation of random values) or by more complex models. The more complex the model is, the more accurate the position estimation will be. However, increasing the model complexity will also increase its cost, the computational requirements, the energy consumption and the time required for the model development [55], [64]. 粒子滤波器（PF）方法是一种数据融合算法，它可以合并不同传感器提供的数据，即合并本体感觉传感器和外部感觉传感器提供的数据。移动节点在给定时刻的位置是根据一组加权粒子计算出来的。粒子是由 PF 生成的位置估计值。基于这种方法的 IPS 需要两个入口模型，即感知模型和运动模型。感知模型根据感知数据确定估计位置（粒子）与设备实际位置匹配的概率。运动模型将粒子运动描述为移动节点运动的近似值。运动可以用简单的模型（如生成随机值）或更复杂的模型来描述。模型越复杂，位置估计就越精确。但是，增加模型的复杂度也会增加成本、计算要求、能耗和模型开发所需的时间 [55]，[64]。
The Pattern Recognition method or Fingerprinting is the localization method used in most IPSs. This method consists of two phases, the offline and the online phase. During the offline phase, a map with the proprieties (fingerprints) of the environment is created - a fingerprint is one characteristic, or a set of characteristics, which makes an environment unique [58]. The fingerprinting map is created by recording the Received Signal Strength Indicator (RSSI) of each anchor node for a specified position in a grid [33]. This process will be repeated for all the points in the grid defined for the site. To minimize the impact of the several interferences that a signal is subjected, various samples are gathered for each position. The final RSSI values are obtained via a weighted average of the collected samples. The denser the grid, the more accurate the IPS will be. However, the complexity and the development time of the system will also increase [12], [31], [58], [64]. During the 模式识别法或指纹识别法是大多数 IPS 使用的定位方法。这种方法包括两个阶段，离线阶段和在线阶段。在离线阶段，创建一张包含环境特征（指纹）的地图--指纹是使环境独一无二的一种特征或一组特征[58]。指纹图是通过记录网格中指定位置每个锚节点的接收信号强度指示器（RSSI）绘制的[33]。这一过程将在为站点定义的网格中的所有点上重复进行。为了尽量减少信号受到的各种干扰的影响，每个位置都要收集不同的样本。最终的 RSSI 值是通过对收集到的样本进行加权平均得到的。网格越密集，IPS 就越精确。不过，系统的复杂性和开发时间也会增加 [12]、[31]、[58]、[64]。在
Fig. 5. Taxonomy of IPSs for emergency responders. 图 5.针对应急响应人员的 IPS 分类。
online phase, the mobile node collects the RSSI values from the various anchor nodes within its range and then compares these values with those stored in the form of a fingerprinting map. The best match is defined as the position of the mobile node. In [31] are summarized some algorithms used for the matching process. The fingerprinting method can be used with RSS and CSI measurements. However, the CSI measurements have shown a tremendous performance gain when compared with RSS-based IPSs [59]-[61]. The major drawback of this method is the need of an extensive site survey. 在线阶段，移动节点从其范围内的各个锚节点收集 RSSI 值，然后将这些值与以指纹图形式存储的值进行比较。最佳匹配值被定义为移动节点的位置。文献 [31] 总结了一些用于匹配过程的算法。指纹识别方法可用于 RSS 和 CSI 测量。不过，与基于 RSS 的 IPS 相比，CSI 测量显示了巨大的性能提升 [59]-[61]。这种方法的主要缺点是需要进行广泛的现场勘测。
The Visual Analysis method determines the position of a mobile node based on images that can be obtained from one or more cameras. This method uses landmarks to estimate the position of a person/object. Since the position of landmarks is known, the position of the mobile nodes is computed by comparison with those points. Based on the scenario, these systems can be classified as static or dynamic. In IPSs based on the visual analysis method, the localization of mobile nodes is passive, i.e., no additional hardware is needed for localization. However, since the localization is carried out based on landmarks, its use is not recommended for unstructured environments [33]. 视觉分析法根据从一个或多个摄像头获取的图像确定移动节点的位置。这种方法利用地标来估计人/物的位置。由于地标的位置是已知的，因此移动节点的位置是通过与这些点的比较计算出来的。根据应用场景，这些系统可分为静态和动态两种。在基于视觉分析方法的 IPS 中，移动节点的定位是被动的，即定位不需要额外的硬件。不过，由于定位是基于地标进行的，因此不建议在非结构化环境中使用[33]。
The Map Matching (MM) method combines electronic maps with the position information (e.g., GPS coordinates and IMU readings) to compute the position of a mobile node. Widely used systems based on this method are the satellite navigation systems. The use of electronic maps is an efficient alternative to hardware installation, as they allow reducing the cost and increasing the scalability of the IPS [43]. However, the 地图匹配（MM）方法将电子地图与位置信息（如 GPS 坐标和 IMU 读数）相结合，计算移动节点的位置。基于这种方法的广泛应用系统是卫星导航系统。使用电子地图是硬件安装的有效替代方法，因为它们可以降低成本，提高 IPS 的可扩展性[43]。但是
availability of electronic maps for unstructured environments is very low and their information may be outdated. 非结构化环境电子地图的可用性非常低，而且其信息可能已经过时。
V. Classification of IPS Schemes for Emergency Responders V.针对应急响应人员的 IPS 方案分类
The technological principle attribute classifies an IPS based on the technologies used for the localization process into four categories: 1) radio signal-based systems; 2) IMU-based systems; 3) hybrid systems and 4) other systems. 技术原理属性根据定位过程中使用的技术将 IPS 分为四类：1) 基于无线电信号的系统；2) 基于 IMU 的系统；3) 混合系统和 4) 其他系统。
Radio Signal-Based Systems: This class of IPSs relies only on wireless communication technologies (e.g., GPS, WiFi, UWB, and ZigBee) to collect the parameters required by the localization method (e.g., RSSI, ToF, AoA, and TDoF). 基于无线电信号的系统：这类 IPS 仅依靠无线通信技术（如 GPS、WiFi、UWB 和 ZigBee）来收集定位方法所需的参数（如 RSSI、ToF、AoA 和 TDoF）。
IMU-Based Systems: This class of IPSs relies on the data acquired from inertial and motion sensors to provide an estimation of the position. Wireless technologies - e.g., RFID and UWB - can also be used in combination with inertial sensors, but unlike radio signal-based system or hybrid systems, these wireless technologies are only used to measure point-to-point distances or to detect the presence of the target node (using the proximity localization method). These 基于 IMU 的系统：这类 IPS 依赖于从惯性传感器和运动传感器获取的数据来估算位置。无线技术（如 RFID 和 UWB）也可与惯性传感器结合使用，但与基于无线电信号的系统或混合系统不同，这些无线技术仅用于测量点对点距离或检测目标节点的存在（使用近距离定位方法）。这些
measurements are just to correct the position error associated with the sensor drift. Another strategy to correct the position estimation is the use of a sitemap. 测量只是为了纠正与传感器漂移相关的位置误差。校正位置估算的另一种策略是使用站点地图。
Hybrid Systems: This class of IPSs relies on both wireless technologies and inertial sensors for the localization of the emergency responders. Each subsystem estimates the position of the emergency responder separately. The final position estimation is computed based on a data fusion algorithm. This algorithm sets different weights to each position estimation, grounded on the estimated error. The main difference between hybrid systems and the IMU-based systems is that in hybrid systems each subsystem is capable of providing a continuous position estimation. 混合系统：这类 IPS 依靠无线技术和惯性传感器对应急响应者进行定位。每个子系统分别估算应急响应者的位置。最终的位置估计是根据数据融合算法计算得出的。该算法根据估计误差为每个位置估计值设置不同的权重。混合系统与基于 IMU 的系统的主要区别在于，在混合系统中，每个子系统都能提供连续的位置估计。
Other Systems: This class represents the IPSs that are based on technologies that do not fit the previous technological principles. Examples of these technologies are the artificial magnetic fields or the “smart” ropes. 其他系统：该类别代表了基于不符合前述技术原则的技术的 IPS。例如人工磁场或 "智能 "绳索。
B. Deployment B.部署
The deployment attribute classifies an IPS according to the placement of the infrastructure needed by an IPS to be operational in the intervention scenario. The deployment is further divided into three categories: 1) pre-deployment; 2) strategic deployment and 3) deployment-free. 部署属性根据 IPS 在干预场景中运行所需的基础设施的位置对 IPS 进行分类。部署又分为三类1) 部署前；2) 战略部署；3) 无部署。
Pre-Deployment: In this class of IPSs, the infrastructure required for the localization of emergency responders is already installed on the building. A typical example of this class of IPSs are the ones that resort the existing wireless communication network. 预部署：在这一类 IPS 中，应急响应人员定位所需的基础设施已经安装在建筑物上。这类 IPS 的典型例子是利用现有无线通信网络的 IPS。
Strategic Deployment: In this class of IPSs, the infrastructure required for the localization of emergency responders is deployed as the emergency responders enter a building. 战略部署：在这一类 IPS 中，当应急响应人员进入建筑物时，就会部署应急响应人员定位所需的基础设施。
Deployment-Free: In this class of IPSs, no infrastructure is required to be deployed. Usually, these IPSs are based on inertial sensors, inertial sensors and map knowledge or GPS. 免部署：这类 IPS 无需部署基础设施。通常，这类 IPS 基于惯性传感器、惯性传感器和地图知识或 GPS。
C. Localization Principle C.定位原理
The localization principle attribute classifies an IPS according to the localization methods used to compute the position of an emergency responder. The localization principle is further divided into five categories: 1) infrastructure-based localization; 2) localization with WSNs; 3) ad-hoc localization; 4) proximity localization and 5) dead reckoning localization. 定位原理属性根据用于计算应急响应者位置的定位方法对 IPS 进行分类。定位原理又分为五类：1) 基于基础设施的定位；2) 使用 WSN 的定位；3) 临时定位；4) 就近定位和 5) 死区定位。
Infrastructure-Based Localization: This class represents the IPSs whose localization method is developed based on the existing infrastructure. Examples of such infrastructure are the wireless communication network (WLAN) or the GPS. A distinct feature of this class is that the developer has no control on the physical specification of the system. So, to improve the performance of the IPS, more robust localization algorithms with higher computational requirements and energy consumption have to be developed [31]. The most used localization method within this class is the fingerprinting (discussed in Section IV-B2). In traditional IPSs applications, the use of infrastructure-based localization is very popular because it avoids the use of expensive infrastructures. 基于基础设施的本地化：这类 IPS 的定位方法是基于现有基础设施开发的。这类基础设施的例子包括无线通信网络（WLAN）或全球定位系统。这类系统的一个显著特点是，开发人员无法控制系统的物理规格。因此，为了提高 IPS 的性能，必须开发计算要求更高、能耗更大的更强大的定位算法[31]。在这类算法中，最常用的定位方法是指纹识别（将在第 IV-B2 节中讨论）。在传统的 IPS 应用中，基于基础设施的定位非常流行，因为它可以避免使用昂贵的基础设施。
Localization With WSNs: This class represents the IPSs that perform the localization of emergency responders based on localization methods developed for WSNs. As previously stated, to perform the localization in WSNs, the nodes have to be densely distributed throughout the building. This means that the nodes have to be pre-deployed in the building or the emergency responders deploy the nodes. Both scenarios are undesirable for an IPS for emergency responders (Section III-C). One advantage of the IPSs based on WSNs is the possibility of monitoring environmental parameters, like temperature, toxic gases and humidity, through the nodes that comprise the WSN [22]. 利用 WSN 定位：这类 IPS 根据为 WSN 开发的定位方法对应急响应人员进行定位。如前所述，要在 WSN 中执行定位，节点必须密集地分布在整个建筑物中。这意味着节点必须预先部署在建筑物中，或者由应急响应人员部署节点。这两种情况都不适合用于应急响应人员的 IPS（第 III-C 节）。基于 WSN 的 IPS 的一个优点是可以通过组成 WSN 的节点监控温度、有毒气体和湿度等环境参数 [22]。
Ad-Hoc Localization: Represents the localization methods that are based on wireless technologies, but the infrastructure is deployed as the emergency responders enter a building. Unlike the infrastructure-based localization, in an ad-hoc localization scheme the designer is responsible for the development of all the signal exchange system and the communication infrastructure. Therefore, the designer has the control over the physical specifications of the IPS, which allows him to control the quality of the localization results. This approach has higher development time and cost. Ad-Hoc 本地化：指基于无线技术的定位方法，但基础设施是在应急响应人员进入建筑物时部署的。与基于基础设施的定位不同，在临时定位方案中，设计者负责开发所有信号交换系统和通信基础设施。因此，设计者可以控制 IPS 的物理规格，从而控制定位结果的质量。这种方法的开发时间和成本较高。
Proximity Localization: The proximity localization method was already discussed in Section IV-B1). In the context of IPSs for emergency responders, this class of localization methods is commonly used to correct the estimated position. Therefore, this method is regarded as a complementary method and is only used for specific situations. 近距离定位：近距离定位方法已在第 IV-B1 节中讨论过。）在用于紧急救援人员的 IPS 系统中，这类定位方法通常用于修正估计位置。因此，这种方法被视为一种补充方法，仅用于特定情况。
Dead Reckoning Localization: This class represents the localization methods that relies on inertial and motion sensors to compute the user position, discussed in Section IV-B1). This class is independent of any infrastructure and can be combined with all the classes discussed before. 惯性定位：该类表示依赖惯性传感器和运动传感器计算用户位置的定位方法（在第 IV-B1 节中讨论）。该类独立于任何基础设施，可与前面讨论的所有类结合使用。
D. Algorithm D.算法
The algorithms developed for the localization process of an IPS can be classified as centralized or decentralized. The main difference between these two classes of algorithms is the platform in which the localization algorithm runs. In a centralized approach, the position estimations of all the target nodes are executed on the same platform, usually, a base station. In a decentralized approach, each target node is responsible for computing its position. The class of the localization algorithm is intrinsically linked with the topology defined for the localization process, discussed in Section IV. So, centralized algorithms are developed for the remote positioning and indirect remote positioning topologies, whereas decentralized algorithms are developed for the self-positioning and indirect self-positioning topologies. 为 IPS 定位过程开发的算法可分为集中式和分散式两种。这两类算法的主要区别在于运行定位算法的平台。在集中式方法中，所有目标节点的位置估计都在同一平台（通常是基站）上执行。在分散式方法中，每个目标节点负责计算自己的位置。定位算法的类别与为定位过程定义的拓扑结构有内在联系，第四节对此进行了讨论。因此，针对远程定位和间接远程定位拓扑结构开发了集中式算法，而针对自定位和间接自定位拓扑结构开发了分散式算法。
E. Environment E.环境
Although the focus of this survey is positioning systems that are capable of localizing emergency responders in indoor environments, outdoor localization is also required. Especially systems that are capable of working in both environments. Therefore, the IPS can also be classified based on the operation environment: indoor, outdoor or both. 虽然本次调查的重点是能够在室内环境中定位应急响应人员的定位系统，但也需要室外定位。特别是能够在这两种环境中工作的系统。因此，IPS 也可根据工作环境进行分类：室内、室外或两种环境。
VI. State-of-the-Art on IPSs for Emergency Responders VI.应急响应人员 IPS 的最新进展情况
Many IPSs have been proposed over the years to solve the problem of localizing emergency responders in indoor environments. Different schemes and strategies have been proposed to improve the IPS performance. In this section, a comprehensive review of the state-of-art IPSs developed specifically for emergency responders is presented. The description of each IPS starts identifying the emergency situation to which it was designed for, followed by a brief description on the system specificities. Then, the working principle of the localization scheme is defined. Finally, each IPS is analyzed based on the requirements discussed in Section III. 多年来，人们提出了许多 IPS，以解决在室内环境中定位应急响应人员的问题。为了提高 IPS 的性能，人们提出了不同的方案和策略。本节将全面回顾专为应急响应人员开发的最先进的 IPS。在对每种 IPS 进行描述时，首先要确定其设计所针对的紧急情况，然后简要介绍系统的具体特点。然后，定义了定位方案的工作原理。最后，根据第三节讨论的要求对每个 IPS 进行分析。
The various IPSs are classified according to the technological principle. Within each subsection, the different IPSs are analyzed in chronological order, to provide the reader a better insight into the evolution of the research on each topic. 各种 IPS 根据技术原理进行分类。在每个小节中，按时间顺序对不同的 IPS 进行分析，以便读者更好地了解每个主题的研究演变情况。
A. Radio Signal-Based Systems A.基于无线电信号的系统
The use of radio frequency (RF) technologies is a widely used approach in the development of IPSs. The radio signalbased systems cover a wide variety of RF technologies, such as: Wi-Fi, RFID, ZigBee , UWB and Bluetooth. 射频（RF）技术的使用是 IPS 开发中广泛使用的一种方法。基于无线电信号的系统涵盖多种射频技术，如：Wi-Fi、RFID、ZigBee、UWB 和蓝牙：Wi-Fi、RFID、ZigBee、UWB 和蓝牙。
A common characteristic of all RF technologies is the ability of radio waves to travel through walls and human bodies. Thus, when compared with other signal-based technologies, like ultrasound and infrared, RF technologies have a larger coverage area, need fewer hardware and in some cases, it is possible to reuse the WLAN infrastructure. In this subsection, the IPSs for emergency responders that rely on RF technologies are presented and discussed. 所有射频技术的共同特点是无线电波能够穿过墙壁和人体。因此，与超声波和红外线等其他基于信号的技术相比，射频技术的覆盖范围更大，所需的硬件更少，在某些情况下还可以重复使用无线局域网基础设施。本小节将介绍和讨论依赖射频技术的应急响应人员 IPS。
SmokeNet is a WSN developed under the Fire Information and Rescue Equipment (FIRE) project to support urban and industrial firefighters [65]. The WSN, previously installed in the building, is designed to perform localization, tracking, environmental monitoring (temperature and presence of smoke) and redundant emergency communication tasks in multi-storey buildings. SmokeNet 是在火灾信息和救援设备（FIRE）项目下开发的一个 WSN，旨在为城市和工业消防员提供支持[65]。该 WSN 先前安装在建筑物中，旨在执行多层建筑中的定位、跟踪、环境监测（温度和烟雾的存在）和冗余应急通信任务。
The system integrates motes installed in each room and near entryways and stairwells- to identify the floor, and motes carried by the firefighters- that can be used as relays for the data exchange in case of destruction of the ones installed in the building. The firefighter’s position is calculated based on RSSI readings and the fingerprinting method. The system provides a room-level accuracy and the position information has a high reliability. All the information collected is made available to the incident commander and the firefighters through an electronic Incident Command System (eICS) software and the FireEye head-mounted display (HMD), respectively. The FireEye HMD is attached to the firefighter’s breathing mask and displays the building layout and small text messages. 该系统集成了安装在每个房间、入口处和楼梯间附近用于识别楼层的微尘和消防员随身携带的微尘，这些微尘可在安装在建筑物内的微尘遭到破坏时用作数据交换的中继器。消防员的位置是根据 RSSI 读数和指纹识别法计算出来的。该系统可提供房间级的精确度，位置信息具有很高的可靠性。收集到的所有信息可分别通过电子事故指挥系统（eICS）软件和 FireEye 头戴式显示器（HMD）提供给事故指挥官和消防员。FireEye HMD 安装在消防员的呼吸面罩上，显示建筑物布局和小文本信息。
The firefighter’s position is updated every two seconds, but if the packet loss increases, the latency of the system will also increase. In real scenarios, the distance for mote to mote communication should be kept at approximately 20 meters. In the case of power outages, the system is powered by batteries 消防员的位置每两秒更新一次，但如果数据包丢失增加，系统的延迟也会增加。在实际场景中，微尘与微尘之间的通信距离应保持在 20 米左右。在停电的情况下，系统由电池供电
and has an autonomy up to 95 hours. This IPS solution is not easily scalable. 其使用寿命长达 95 小时。这种 IPS 解决方案不易扩展。
The Europcom is an indoor positioning and communication system developed to assist emergency personnel in rescue operations in disaster zones [17], [66], [67]. This system is an ad-hoc network comprising four different elements, Control Unit (CU), Base Units (BU), Mobile Unit (MU) and Distributed Units (DU). Each component has a different function in the network: the CU is a Graphical User Interface (GUI) developed to display the positions of the emergency personnel to the incident commander; the BUs are attached to the emergency vehicles and have referencing and communications functions; the MUs are responsible for calculating the user’s position; and the DUs are designed to expand the network range and increase the position accuracy. For localization and communication, this system uses the UWB and Wi-Fi technologies, respectively. The position of the emergency personnel is calculated based on RToF measurements and the iterative least square algorithm, with an accuracy of 1-21-2 meters. The emergency personnel position information is displayed to the incident commander through the CU GUI. Europcom 是一个室内定位和通信系统，用于协助应急人员在灾区开展救援行动 [17]、[66]、[67]。该系统是一个由控制单元（CU）、基地单元（BU）、移动单元（MU）和分布式单元（DU）四个不同部分组成的特设网络。每个组件在网络中都有不同的功能：CU 是一个图形用户界面 (GUI)，用于向事故指挥官显示应急人员的位置；BU 安装在应急车辆上，具有参考和通信功能；MU 负责计算用户的位置；DU 用于扩大网络范围和提高位置精度。在定位和通信方面，该系统分别采用了 UWB 和 Wi-Fi 技术。应急人员的位置根据 RToF 测量和迭代最小平方算法计算得出，精度为 1-21-2 米。应急人员的位置信息通过 CU GUI 显示给事故指挥官。
The system has to be deployed by the emergency personnel, therefore, is independent of the building infrastructure and prior data collection. DUs can be deployed to expand the coverage of the system. The dimensions of the MUs are higher than the maximum volume defined by emergency responders (107.34cm^(3))\left(107.34 \mathrm{~cm}^{3}\right), as stated in Section III-D. 该系统必须由应急人员部署，因此不受建筑基础设施和事先数据收集的影响。可以部署单元来扩大系统的覆盖范围。如第 III-D 节所述，MU 的尺寸高于应急响应人员定义的最大容积 (107.34cm^(3))\left(107.34 \mathrm{~cm}^{3}\right) 。
The ProeTEX project was designed to monitor physiological, activity and environmental related parameters of emergency responders, more specifically Civil Protection, Urban and Forest firefighters [50], [68]. These parameters are collected and processed through a Body Area Network (BAN) and sent in real-time to the command post through a deployed remote transmission system based on the Wi-Fi technology [68]. The communication network has a mesh topology. The position of each emergency responder is obtained using a GPS receiver built-in in the BAN. As the position estimation in indoor scenarios is unreliable with the GPS, the system only provides a symbolic representation of the emergency responders in the building. All the information is collected with a frequency of 1 Hz and displayed through a GUI at the command post side. ProeTEX 项目旨在监测应急响应人员，特别是民防、城市和森林消防员的生理、活动和环境相关参数[50]、[68]。这些参数通过体域网（BAN）收集和处理，并通过基于 Wi-Fi 技术的远程传输系统实时发送到指挥所[68]。通信网络采用网状拓扑结构。每个应急响应者的位置通过 BAN 内置的 GPS 接收机获得。由于全球定位系统在室内场景下的位置估计并不可靠，因此该系统只能提供建筑物内应急响应人员的符号表示。所有信息均以 1 Hz 的频率收集，并通过指挥所端的图形用户界面显示出来。
The localization method is deployment-free, but the system still requires the deployment of the communication network. Inside buildings, the system has a maximum coverage of 156 meters. For ease of assembling and to increase wearer’s comfort, a modular system is proposed, with some sensors embedded on the garment, resulting in small size and lightweight modules, validated under harsh environments [69]. The modules are powered by a single battery that guarantees an autonomy up to 7 hours. The different modules are wire connected which decreases the scalability of the system. 这种定位方法无需部署，但系统仍需要部署通信网络。在建筑物内，该系统的最大覆盖范围为 156 米。为了便于装配和提高佩戴者的舒适度，我们提出了一种模块化系统，将一些传感器嵌入服装中，从而形成了体积小、重量轻的模块，并在恶劣的环境中得到了验证[69]。这些模块由单个电池供电，可保证自主运行长达 7 小时。不同的模块通过电线连接，降低了系统的可扩展性。
The LifeNet system was designed to support firefighters during interventions [70]. This system is inspired by the traditional search systems based on ropes and, similarly, aims to provide assisted navigation functionality to quickly and easily find the nearest exit. LifeNet 系统的设计目的是在干预过程中为消防员提供支持[70]。该系统受到传统绳索搜索系统的启发，同样旨在提供辅助导航功能，以便快速轻松地找到最近的出口。
The system relies on several sensor nodes, carried by a firefighter in a device called “Beacon Ejector”, which are automatically deployed based on the RSS from other sensor nodes already deployed. The sensor nodes form an ad-hoc network creating virtual paths, where each sensor node acts as a waypoint to guide the firefighter. These virtual paths can be used to create escape routes and shortcuts. 该系统依靠消防员携带的名为 "信标弹射器 "的装置中的多个传感器节点，这些节点根据已部署的其他传感器节点的 RSS 自动部署。这些传感器节点组成一个特设网络，创建虚拟路径，其中每个传感器节点都是引导消防员前进的路标。这些虚拟路径可用于创建逃生路线和捷径。
This system only provides a relative position information of the firefighter, which is displayed in an HMD integrated in the breathing mask. The coverage of the system is limited to the number of sensor nodes transported in each beacon ejector module and to the measurement range of the sensor nodes ( 2.5 meters in the current implementation). In the current version, some components of the system are too bulky to be used by firefighters (e.g., the wearable computer). The sensor nodes are battery powered, but the information about their autonomy is not provided. 该系统仅提供消防员的相对位置信息，并通过集成在呼吸面罩中的 HMD 显示。该系统的覆盖范围受限于每个信标弹射器模块中的传感器节点数量和传感器节点的测量范围（目前的实施范围为 2.5 米）。在当前版本中，系统的某些组件过于笨重，消防员无法使用（如可穿戴计算机）。传感器节点由电池供电，但没有提供有关其自主性的信息。
The FIREGUIDE is a localization and navigation system developed to assist firefighters during interventions [71]. This system aims to guide the firefighters to the nearest exit and shows the incident commander their position on the building map. For the localization, this IPS relies on RFID tags (R-tags) and Bluetooth tags (B-tags) preinstalled on the building and on a PDA equipped with a RFID reader. The R-tags provide a high accuracy localization ( 2-3cm2-3 \mathrm{~cm} ) and the B-tags provide a room-level accuracy. Two display units were developed, one runs in a PDA and aims to guide the firefighter to the nearest exit, while the other runs on a server and displays all the positions of the firefighters to the incident commander. FIREGUIDE 是一种定位导航系统，用于协助消防员进行干预[71]。该系统旨在引导消防员找到最近的出口，并向事故指挥官显示他们在建筑地图上的位置。该 IPS 依靠预装在建筑物上的 RFID 标签（R-tags）和蓝牙标签（B-tags）以及配备 RFID 阅读器的 PDA 进行定位。R 标签提供高精度定位（ 2-3cm2-3 \mathrm{~cm} ），B 标签提供房间级精度。开发了两个显示单元，一个在 PDA 上运行，旨在引导消防员找到最近的出口，另一个在服务器上运行，向事故指挥官显示消防员的所有位置。
The system requires an offline phase for the deployment and mapping of the R- and B-tags and the creation of the building map. To communicate with the incident commander this IPS uses the existing WLAN network. Therefore, it is highly dependent on the building infrastructure and its coverage is limited by the tags deployed beforehand. Moreover, the system is not easily scalable. The R- and B-tags rely on the building power and, in a case of a breakdown of the grid, the system will not work. 该系统需要一个离线阶段来部署和绘制 R 标记和 B 标记，并创建建筑地图。为了与事件指挥员进行通信，IPS 使用现有的 WLAN 网络。因此，它高度依赖于楼宇基础设施，其覆盖范围也受到事先部署的标签的限制。此外，该系统不易扩展。R 和 B 标签依赖于楼宇电力，一旦电网发生故障，系统将无法工作。
Rüppel et al. [72] proposed a system to support rescuers in finding the shortest way within complex buildings. This IPS is based on a framework that follows a Multi-Method-Approach (MMA), i.e., the combination of different positioning technologies (Wi-Fi, UWB and RFID) to achieve different granularities, in terms of position accuracy and a larger coverage. Each positioning technology is based on COTS systems, namely the Ekahau [4], the Ubisense [2] and the Identec Solutions [73], for Wi-Fi, UWB and RFID, respectively. Based on the positioning technology used, different position accuracies will be obtained: RFID provides a roomlevel accuracy, UWB a decimeter-level position accuracy and Wi-Fi a meter-level position accuracy. The position information of the rescue units is combined with the digital route cards and displayed in portable Tablet PCs. The use of Tablet PCs is not suitable for firefighters, because its handling interferes with their on-mission tasks. Rüppel 等人[72]提出了一种系统，用于支持救援人员在复杂的建筑物内寻找最短路径。该 IPS 基于一个多方法（MMA）框架，即结合不同的定位技术（Wi-Fi、UWB 和 RFID），以实现不同的定位精度和更大的覆盖范围。每种定位技术都基于 COTS 系统，即分别用于 Wi-Fi、UWB 和 RFID 的 Ekahau [4]、Ubisense [2] 和 Identec Solutions [73]。根据所使用的定位技术，将获得不同的定位精度：RFID 提供房间级精度，UWB 提供分米级定位精度，Wi-Fi 提供米级定位精度。救援单元的位置信息与数字路线卡相结合，并显示在便携式平板电脑上。平板电脑不适合消防员使用，因为操作平板电脑会影响他们执行任务。
All the systems of this framework have to be preinstalled on the building and require the collection of data beforehand. The coverage of the system depends on the position 该框架的所有系统都必须预先安装在建筑物上，并需要事先收集数据。系统的覆盖范围取决于位置
technology used, e.g., Wi-Fi has a higher coverage than UWB, but a lower accuracy. All the COTS IPSs depend on the building power. 例如，Wi-Fi 的覆盖率比 UWB 高，但精度较低。所有 COTS IPS 都依赖于楼宇电源。
Li and Becerik-Gerber [20] proposed an IPS for emergency response operations that is capable of locating the building occupants and first responders. The system relies on smartphones and in the use of tethering technology to establish an ad-hoc Wi-Fi network among the smartphones and to collect the parameters of interest (SSID, RSSI, and MAC address). The SSID and the MAC address identify a smartphone and the RSSI is used to calculate the distance between devices. Based on the distances calculated, the trilateration technique is used to locate the firefighters and occupants. Li 和 Becerik-Gerber[20]提出了一种用于应急响应行动的 IPS，该系统能够定位楼宇住户和急救人员。该系统依靠智能手机和系绳技术在智能手机之间建立一个特设 Wi-Fi 网络，并收集相关参数（SSID、RSSI 和 MAC 地址）。SSID 和 MAC 地址用于识别智能手机，RSSI 用于计算设备之间的距离。根据计算出的距离，使用三坐标技术确定消防员和住户的位置。
The mean accuracy of the system is 2.24 meters, evaluated in a room scale test, in LOS conditions. The localization algorithm takes 8-12 seconds to compute the positions of firefighters and occupants, but to convert the RSSI values into distances, specific building parameters have to be estimated beforehand. This IPS has two modes of operation, normal and recovery. The normal mode relies on the routers of the existing WLAN infrastructure, whose position is previously known. The recovery mode is infrastructure-free and is designed for scenarios of power outages. In both modes, the system must know two absolute positions (from the routers or firefighters) to calculate the positions of the other firefighters and occupants. How the position information is presented is not mentioned. 系统的平均精确度为 2.24 米，这是在 "LOS "条件下进行的房间规模测试中评估得出的结果。定位算法只需 8-12 秒即可计算出消防员和住户的位置，但要将 RSSI 值转换为距离，必须事先估算出具体的建筑参数。该 IPS 有两种运行模式，即正常模式和恢复模式。正常模式依赖于现有无线局域网基础设施的路由器，而路由器的位置是已知的。恢复模式不依赖基础设施，专为断电情况设计。在这两种模式下，系统必须知道两个绝对位置（来自路由器或消防员），才能计算出其他消防员和住户的位置。至于如何显示位置信息，系统并未提及。
The coverage of the system is limited by the short communication range of the smartphones ( ~~30\approx 30 meters) and by the fact that each device has to be connected to other two devices to run the localization algorithm. No scalability tests were performed, but, in areas with a high density of devices, more combinations of connections between devices are possible, which will increase the latency of the localization algorithm and may affect the requirement of real-time calculation of the position. The autonomy of the system depends on the battery of the smartphones. 该系统的覆盖范围受到智能手机通信距离短（ ~~30\approx 30 米）以及每个设备必须与其他两个设备连接才能运行定位算法的限制。虽然没有进行可扩展性测试，但在设备密度较高的地区，设备之间可能会有更多的连接组合，这将增加定位算法的延迟，并可能影响实时计算位置的要求。系统的自主性取决于智能手机的电池。
Moon et al. [74], [75] proposed an IPS to support firefighters in a operation scenario. The system comprises three elements: 1) a monitoring station, placed outside the building and equipped with a GPS receiver to obtain the absolute position information; 2) portable access points (APs), deployed at the entrance of the building for the initialization process and distributed as the firefighters enter a building to expand the IPS coverage; and 3) the firefighters, equipped with an AP. Each AP consists of two ranging devices, one acts as an anchor and the other as a tag, and both are based on the UWB technology. The localization process is carried out in two stages [75]. In the first stage, each tag collects the ranging measurements from all the hearable anchors. Based on the ranging measurements and the EKF and SLAM methods, each tag calculates its relative initial position and sends all the information to the monitoring station through the ad-hoc network. In the second stage, the monitoring station corrects the initial position estimations of the APs and the firefighters by recalculating their positions. Moon 等人[74]、[75] 提出了一种 IPS，用于在行动场景中为消防员提供支持。该系统由三部分组成：1) 监控站，放置在建筑物外，配备 GPS 接收器，用于获取绝对位置信息；2) 便携式接入点（AP），部署在建筑物入口处，用于初始化过程，并在消防员进入建筑物时分布，以扩大 IPS 的覆盖范围；3) 消防员，配备一个接入点。每个 AP 由两个测距设备组成，一个作为锚点，另一个作为标签，两者都基于 UWB 技术。定位过程分两个阶段进行 [75]。在第一阶段，每个标签从所有可听到的锚点收集测距数据。根据测距结果和 EKF 及 SLAM 方法，每个标签计算出自己的相对初始位置，并通过 ad-hoc 网络将所有信息发送到监测站。在第二阶段，监测站通过重新计算 AP 和消防员的位置来修正其初始位置估计值。
The position error reported for the IPS lays below 2 meters, but the tests performed are not enough to assess the real 据报告，IPS 的位置误差低于 2 米，但所进行的测试不足以评估其实际误差。
performance of the system in a real scenario. How the position information is presented is not mentioned. The system is capable of expanding its coverage by deploying more APs. An order to deploy a new AP is triggered every time the number of detectable APs is less than three or the error covariance is larger than a predefined threshold. In the current stage, this IPS is only capable of localizing one firefighter. The information on IPS autonomy is not provided. 系统在实际场景中的性能。至于如何呈现位置信息，则没有提及。该系统能够通过部署更多的接入点来扩大覆盖范围。每当可探测到的 AP 数量少于 3 个或误差协方差大于预定阈值时，就会触发部署新 AP 的命令。在现阶段，该 IPS 只能定位一名消防员。没有提供有关 IPS 自主性的信息。
Ghosh et al. [76] proposed a system for localization, tracking and monitoring different physiological (e.g., heart rate, body temperature and pulse) and environmental ( CO and HCN ) parameters during real firefighting scenarios. This IPS is based on an ad-hoc network integrating sensor motes deployed as the firefighters enter a building and a base station placed outside the building. The sensor motes are ZigBee compliant and based on the TelosB platform developed by Texas Instruments. The localization of the firefighters is performed by monitoring the drop in the RSSI values when a firefighter obstructs the communication between two sensor motes. Ghosh 等人[76] 提出了一种系统，用于在真实消防场景中定位、跟踪和监测不同的生理（如心率、体温和脉搏）和环境（一氧化碳和六氯苯）参数。该 IPS 基于一个 ad-hoc 网络，该网络整合了消防员进入建筑物时部署的传感器微尘和建筑物外的基站。传感器微尘符合 ZigBee 标准，基于德州仪器公司开发的 TelosB 平台。消防员的定位是通过监测当消防员阻碍两个传感器马达之间的通信时 RSSI 值的下降来实现的。
In the proposed IPS scheme, the sensor motes transmit data packets at every 250 milliseconds and the accuracy of the localization can be improved increasing the packets exchange rate. Some preliminary studies have been performed to validate the feasibility of the system, but the accuracy is not provided. Moreover, false positive detections may occur due to the dynamic changes in the environment or even fluctuation of RSSI values due to the multipath phenomena. A GUI is used at the base station to indicate to the incident commander the position of firefighters in the building, but no details are given. Since the IPS is based on an ad-hoc network, the coverage can be expanded if the firefighters have more nodes to deploy when entering a building. The sensor motes are battery powered and have an autonomy up to two months. To increase the battery life, power saving features were incorporated on the sensor motes, namely, they are put in a sleep mode when not sensing or transmitting. 在拟议的 IPS 方案中，传感器马达每 250 毫秒发送一次数据包，提高数据包交换率可提高定位精度。为了验证系统的可行性，我们进行了一些初步研究，但并未提供准确性。此外，由于环境的动态变化，甚至由于多径现象造成的 RSSI 值波动，可能会出现假阳性检测。基站使用图形用户界面向事故指挥员显示消防员在建筑物中的位置，但没有提供详细信息。由于 IPS 基于 ad-hoc 网络，如果消防员在进入建筑物时有更多的节点可以部署，则覆盖范围可以扩大。传感器微尘由电池供电，使用寿命长达两个月。为了延长电池寿命，传感器微尘具有省电功能，即在不传感或不传输时处于睡眠模式。
Zhang et al. [53] proposed an IPS to support emergency rescue operations in critical complex environments. This IPS is based on a self-configured UWB ad-hoc network comprising UWB readers, one UWB central controller and UWB tags. The UWB central controller is installed on an emergency vehicle and is responsible for automatically organize the ad-hoc network and calculate the firefighters’ positions. The firefighters, unmanned air vehicles (UAVs) and robots are equipped with UWB tags that broadcast four signal beams per second. These UWB signal beams carry the transmitting timestamp and are used in the calculation of the TDoA, performed by the UWB readers that are deployed around the building. These measurements are sent to the UWB central controller to calculate the position of the firefighter based on trilateration methods. An EKF is applied to deal with uncertainty factors associated with the wireless links and the stochastic sensor scheduling. The different sources of uncertainty are modeled with different Hidden Markov Model (HMM) components that are integrated into a multi-dimensional Modular Markov Jump System (MMJS). Zhang 等人[53]提出了一种 IPS，用于支持关键复杂环境中的紧急救援行动。该 IPS 基于一个由 UWB 阅读器、一个 UWB 中央控制器和 UWB 标签组成的自配置 UWB ad-hoc 网络。UWB 中央控制器安装在应急车辆上，负责自动组织特设网络并计算消防员的位置。消防员、无人驾驶飞行器（UAV）和机器人都配备了 UWB 标签，每秒发射四束信号。这些 UWB 信号波束带有发射时间戳，用于计算 TDoA，由部署在大楼周围的 UWB 阅读器执行。这些测量结果被发送到 UWB 中央控制器，以根据三坐标法计算消防员的位置。EKF 用于处理与无线链路和随机传感器调度相关的不确定性因素。不同的不确定性来源采用不同的隐马尔可夫模型（HMM）组件建模，这些组件被集成到一个多维模块化马尔可夫跃迁系统（MMJS）中。
In LOS conditions, the firefighter’s position was computed with high accuracy - an error of 0.03 meters of the total 在 LOS 条件下，消防员位置的计算精度很高，总误差为 0.03 米。
distance traveled-, but in NLOS conditions, the IPS was not able to provide a position estimation. Nevertheless, the system gives feedback about the connectivity of a tag to the UWB readers and allows the operation support personnel to monitor and fine tune operations. To ensure information security, authentication is required and the messages are encrypted. The IPS is independent of the building infrastructure, but the UWB readers are connected to the UWB central controller by cables. Handling cables during an emergency scenario can be time-consuming, causing an undesirable deployment delay. To expand the coverage, the firefighters carry portable UWB readers to deploy when the number of detectable UWB readers is lower than three. The IPS is designed based on a plug-in architecture to facilitate the adoption of new localization and networking algorithms and the integration with other systems. No considerations are made on system’s autonomy. 但是，在 NLOS 条件下，IPS 无法提供位置估计。不过，该系统可以反馈标签与 UWB 阅读器的连接情况，并允许操作支持人员监控和微调操作。为确保信息安全，系统需要进行身份验证，并对信息进行加密。IPS 独立于楼宇基础设施，但 UWB 阅读器通过电缆连接到 UWB 中央控制器。在紧急情况下处理电缆可能会很耗时，从而造成不必要的部署延迟。为了扩大覆盖范围，消防员携带便携式 UWB 阅读器，以便在可探测到的 UWB 阅读器数量少于三个时进行部署。IPS 采用插件式结构设计，便于采用新的定位和联网算法，并与其他系统集成。没有考虑系统的自主性。
Li et al. [51], [77], [78] proposed a framework for indoor localization of the first responders in fire emergency response operations. The framework is centered on the Building Information Modeling (BIM) and comprises three major components: RF transmitters, smartphones and a localization server. The RF transmitters are COTS programmable routers and are used to broadcast an RF signal with the RF transmitter unique SSID and MAC address. The smartphones are the mobile units that scan all the detectable Wi-Fi RF transmitters and send the collected information to the remote localization server. The remote localization server computes the first responders’ position based on the received data and displays it on a BIM platform that acts as a GUI. Li 等人[51]、[77]、[78] 提出了一个用于火灾应急响应行动中急救人员室内定位的框架。该框架以建筑信息建模（BIM）为中心，由三个主要部分组成：射频发射器、智能手机和定位服务器。射频发射器是 COTS 可编程路由器，用于广播带有射频发射器唯一 SSID 和 MAC 地址的射频信号。智能手机是扫描所有可探测到的 Wi-Fi 射频发射器的移动装置，并将收集到的信息发送到远程定位服务器。远程定位服务器根据接收到的数据计算急救人员的位置，并将其显示在作为图形用户界面的 BIM 平台上。
Two different approaches were designed to tackle the indoor localization problem. On the one hand, the Iterative Maximum Likelihood Estimation (IMLE) algorithm, designed for scenarios where the existing infrastructure can be accessed to collect the RF data used in the localization process [78]. This algorithm relies on the MLE method to estimate the parameter values of the propagation model created for the RF signal. The combination of the propagation model, the collected RF signals and the MLE method allows computing the first responder’s location. On the other hand, the EnvironmentAware radio frequency beacon deployment algorithm for Sequence-Based Localization (EASBL), designed for scenarios where the infrastructure is inaccessible or inexistent and an ad-hoc network has to be deployed at the emergency site [51]. The EASBL algorithm, which is based on Sequence Based Localization (SBL) [79] has a dual-objective function that balances the localization accuracy and the deployment effort. The optimal solution is computed through the metaheuristic technique Tabu search. The SBL algorithm divides the 2D space into regions, through perpendicular bisectors created by the distance between each pair of anchor nodes. Each region is unique and the first responder position is computed based on the collected RSSI values and the centroid method. 为解决室内定位问题，设计了两种不同的方法。一方面是迭代最大似然估计（IMLE）算法，该算法专为可利用现有基础设施收集定位过程中使用的射频数据的场景而设计[78]。该算法依靠 MLE 方法来估计为射频信号创建的传播模型的参数值。将传播模型、收集到的射频信号和 MLE 方法结合起来，就能计算出第一响应者的位置。另一方面，基于序列定位的环境感知射频信标部署算法（EASBL）设计用于基础设施无法进入或不存在的场景，必须在应急现场部署一个 ad-hoc 网络[51]。EASBL 算法以基于序列的定位（SBL）[79] 为基础，具有平衡定位精度和部署工作量的双目标功能。通过元启发式技术 Tabu 搜索计算出最优解。SBL 算法通过每对锚节点之间的距离创建的垂直平分线，将二维空间划分为若干区域。每个区域都是唯一的，并根据收集到的 RSSI 值和中心点方法计算第一响应者的位置。
Both algorithms are capable of providing room-level and coordinate-level position information. The reported room-level accuracy was above 82.8%82.8 \% and the coordinate-level accuracy above 2.29 meters, in 95%95 \% of the measurements. The IMLE algorithm showed better performance than the EASBL algorithm. The position of the first responders can be visualized by 这两种算法都能提供房间级和坐标级位置信息。在 95%95 \% 的测量中，报告的房间级精度高于 82.8%82.8 \% ，坐标级精度高于 2.29 米。IMLE 算法的性能优于 EASBL 算法。第一反应者的位置可通过以下方式直观显示
the incident commander on the BIM platform, but this information is not available to the first responders. The position of each first responder is updated every five seconds. 但第一响应人员无法获得这些信息。每位急救人员的位置每五秒更新一次。
Both algorithms require site-specific parameters for the construction of the BIM. The IMLE algorithm requires the RF transmitters already installed on the building. With the EASBL algorithm the deployment and set up of the entire network took 90 seconds, which is less than the 135 seconds discussed in Section III-C. 这两种算法都需要特定场地的参数来构建 BIM。IMLE 算法需要在建筑物上安装射频发射器。使用 EASBL 算法，整个网络的部署和设置耗时 90 秒，少于第 III-C 节中讨论的 135 秒。
Both algorithms were designed to cover the entire building, but the solution is site-specific. The IMLE algorithm showed a better robustness when some of the RF transmitters were destroyed. The total size and weight of the IPS components (RF transmitter plus smartphone) comply the requirements defined in Section III-D. The smartphones are battery powered, but no considerations on the system’s autonomy are made. 这两种算法都是为了覆盖整栋大楼而设计的，但解决方案是因地制宜的。当一些射频发射器被破坏时，IMLE 算法显示出更好的鲁棒性。IPS 组件（射频发射器和智能手机）的总尺寸和重量符合第 III-D 节中定义的要求。智能手机由电池供电，但不考虑系统的自主性。
Femminella and Reali [18] proposed an IPS that is capable of tracking first responders in emergency scenarios. This system resorts from the GPS and Wi-Fi technologies for the localization and comprises three major components: a client, several APs, deployed around the building in areas with good visibility, and a localization server. The client is typically a smartphone, equipped with a Wi-Fi module and a GPS receiver, which collects the data from the deployed APs and calculates its position based on the received data. The APs are equipped with a GPS receiver and are configured to write their position in the SSID of the beacon message that is periodically broadcasted. The localization server manages the network communication, storing the current and previous estimated positions and displaying the first responders’ position. Femminella 和 Reali [18] 提出了一种能够在紧急情况下跟踪急救人员的 IPS。该系统利用 GPS 和 Wi-Fi 技术进行定位，由三个主要部分组成：一个客户端、几个部署在建筑物周围能见度高的区域的接入点和一个定位服务器。客户端通常是一部智能手机，配备一个 Wi-Fi 模块和一个 GPS 接收机，负责收集所部署 AP 的数据，并根据接收到的数据计算自己的位置。接入点配有 GPS 接收器，并被配置为将其位置写入定期广播的信标信息的 SSID 中。定位服务器管理网络通信，存储当前和以前的估计位置，并显示第一响应者的位置。
This IPS is based on the concept of pseudolites, i.e., the creation of virtual satellites to increase the coverage and performance of traditional GPS-based systems in areas with poor visibility. Once the position of each AP is set, it starts to broadcast the AP’s position in the SSID information element, acting as a pseudolite. The position of the client device is computed locally and can be performed in one of the following ways: 1) standard GPS, in open areas; 2) an hybrid scheme of IEEE 802.11 and GPS, in partially obstructed outdoor areas; and 3) fully IEEE 802.11, in indoor areas, based on the RSSI values and the trilateration method. 这种 IPS 基于伪卫星概念，即创建虚拟卫星，以提高传统 GPS 系统在能见度较低地区的覆盖范围和性能。一旦设定了每个接入点的位置，它就会开始在 SSID 信息元素中广播接入点的位置，起到伪卫星的作用。客户端设备的位置在本地计算，可以通过以下方式之一进行：1) 开放区域的标准 GPS；2) 部分受阻的室外区域的 IEEE 802.11 和 GPS 混合方案；3) 室内区域的完全 IEEE 802.11，基于 RSSI 值和三角测量法。
The reported accuracy of the system is around 10 meters, which does not meet the requirements discussed in Section III-A. A cartographic client is used to display the first responders’ position to the incident commander and is updated every 10 seconds. The IPS is independent on the building infrastructure or prior data collection. To increase the coverage, the use of drones or robots equipped with an AP to act as pseudolites was proposed, but not implemented. Therefore the coverage of the IPS is limited by the propagation range of Wi-Fi signals in a building. The system can track up to 800 devices, which is in line with the requirements discussed in Section III-D. Although the client device is battery powered, no considerations about the power source of the APs and the autonomy of the IPS are made. 据报告，该系统的精确度约为 10 米，不符合第 III-A 节中讨论的要求。制图客户端用于向事故指挥官显示急救人员的位置，每 10 秒钟更新一次。IPS 不受建筑基础设施或事先数据收集的影响。为了扩大覆盖范围，有人提议使用配备 AP 的无人机或机器人充当伪卫星，但未付诸实施。因此，IPS 的覆盖范围受到建筑物内 Wi-Fi 信号传播范围的限制。该系统最多可跟踪 800 台设备，符合第 III-D 节中讨论的要求。虽然客户端设备由电池供电，但没有考虑 AP 的电源和 IPS 的自主性。
Summary of Radio Signal-based Systems: The use of RF technologies in the development of an IPS has many advantages, namely: (1) the ability of the radio waves to 基于无线电信号的系统概述：在开发 IPS 时使用射频技术有许多优点，即：(1) 无线电波能
travel through walls and human bodies; (2) the accuracy of the system is not affected by the movement pattern of the emergency responder; (3) the possibility to increase the performance of the system by deploying anchor nodes as the emergency responders enter a building; (4) the reuse of the positioning infrastructure for the communication with the incident commander, and (5) the variety of technologies available and the different granularities, in terms of accuracy provided by each technology. However, in complex indoor environments, RF signals are strongly affected by the radio propagation phenomena, which are difficult to predict due to diversity in number and materials of walls on the signal path. Moreover, in emergency scenarios, the combination of high temperatures, falling debris, thick smoke, noise, obstacles and gusts of air, hinder the propagation of RF signals [22]. (2) 系统的准确性不受应急响应者移动模式的影响；(3) 可以在应急响应者进入建筑物时部署锚节点，从而提高系统性能；(4) 可重复使用定位基础设施与事故指挥官进行通信；(5) 可利用的技术种类繁多，且每种技术提供的准确性粒度不同。然而，在复杂的室内环境中，射频信号会受到无线电传播现象的强烈影响，而由于信号路径上墙壁的数量和材料的多样性，这种现象很难预测。此外，在紧急情况下，高温、碎片坠落、浓烟、噪音、障碍物和阵风等因素都会阻碍射频信号的传播[22]。
The localization methods used in radio signal-based systems benefit from the number and spatial distribution of the anchor nodes. Usually, a radio signal-based IPS requires a minimum of three detectable anchor nodes to compute a position estimation, and a correct spatial distribution of the anchor nodes results in a higher accuracy. In a real mission scenario, the availability of three detectable anchor nodes to continuously compute the positions of the emergency responders is improbable and the ideal spatial distribution of the anchor nodes may not be feasible due, on one hand to the geometry restrictions of the intervention scenario and on the other hand, to the short time available for the deployment stage. All these factors combined have a negative impact on the performance of radio signal-based IPSs in terms of accuracy, reliability of the position information and coverage. 基于无线电信号的系统中使用的定位方法得益于锚节点的数量和空间分布。通常情况下，基于无线电信号的 IPS 至少需要三个可探测的锚节点来计算位置估计值，而正确的锚节点空间分布会带来更高的精度。在实际任务场景中，一方面由于干预场景的几何限制，另一方面由于部署阶段的可用时间很短，因此不可能有三个可探测的锚节点来连续计算应急响应人员的位置，而且理想的锚节点空间分布可能也不可行。所有这些因素加在一起，对基于无线电信号的 IPS 在准确性、位置信息可靠性和覆盖范围方面的性能产生了负面影响。
B. IMU-Based Systems B.基于 IMU 的系统
During the past decade, the inertial and motion sensors were the target of a major development. The research community focused on size reduction, integration of several sensors (3D accelerometers, gyroscopes and magnetometers) in a single chip and improvement of their performance through the bias and drift reduction/control. This led to increasingly smaller IMUs with better performance and easy assembly in wearable devices. 过去十年间，惯性传感器和运动传感器成为重大发展的目标。研究界的重点是缩小尺寸、在单个芯片中集成多个传感器（三维加速度计、陀螺仪和磁力计），以及通过减少偏差和漂移/控制来提高其性能。这使得 IMU 的体积越来越小，性能越来越好，并且易于装配到可穿戴设备中。
In this section, the IPSs for emergency responders centered on the IMU information for the localization process are described and discussed. 本节将介绍和讨论以用于定位过程的 IMU 信息为中心的应急响应 IPS。
Beauregard [80] proposed an IPS for public safety and military applications based on a helmet-mounted IMU and a GPS receiver. The GPS is used for outdoor localization while the IMU is used to compute the wearer’s position in GPSdenied environments. The localization process of the PDR technique is decomposed into three steps: a) in the step detection phase, the magnitude of the acceleration is calculated from the accelerometer signals triad and the step boundaries defined by the positive-going zero crossings of a low-pass filtered version of this signal; b) in the numerical step, the features used in a feedforward neural network as training patterns are calculated. The outputs of the neural network are the step lengths estimated from GPS position fixes, interpolated to footfall occurrences and c) the heading estimation is based Beauregard [80] 提出了一种基于头盔安装的 IMU 和 GPS 接收器的公共安全和军事应用 IPS。GPS 用于室外定位，而 IMU 则用于计算佩戴者在无 GPS 环境中的位置。PDR 技术的定位过程分为三个步骤：a) 在台阶检测阶段，根据加速度计信号三元组计算加速度大小，并根据该信号低通滤波版本的正向零交叉定义台阶边界；b) 在数值步骤中，计算前馈神经网络中用作训练模式的特征。神经网络的输出是根据 GPS 定位估算出的步长，并根据脚步发生率进行内插。
on the magnetometer yaw output. To achieve a reliable estimation, the IMU has to be in a fixed orientation relatively to the wearer’s body and the helmet must be pointed towards the direction of the movement. 磁力计的偏航输出。要获得可靠的估算结果，IMU 必须处于相对于佩戴者身体的固定方位，头盔必须指向运动方向。
The position offset reported after a 30 -minute walk is about 40 meters. In real emergency scenarios, the error is likely to be higher since the IPS cannot distinguish the gaze orientation from the direction of travel. Since the localization algorithm runs offline, neither the incident commander nor the emergency responders have access to the position information in real-time. No information on the system’s autonomy is given. 步行 30 分钟后报告的位置偏移约为 40 米。在实际紧急情况下，误差可能会更大，因为 IPS 无法区分注视方向和行进方向。由于定位算法是离线运行的，因此事故指挥官和应急响应人员都无法实时获取位置信息。没有提供关于系统自主性的信息。
Beauregard [81] proposed a different approach for the localization of first responders in typical search and rescue missions. The IMU is mounted on the shoe, which allows the calculation of the displacement of the feet between footfalls directly through inertial mechanization equations. Besides the standard strap-down mechanization equations, this IPS also implements the ZUPT technique to prevent the exponential growth of the error, which is typical in IMU-based IPSs. The combination of both techniques allows reducing and minimizing the inherent drift of IMUs. Beauregard [81] 提出了一种不同的方法，用于在典型的搜救任务中对急救人员进行定位。IMU 安装在鞋子上，可以直接通过惯性机械化方程计算脚步之间的位移。除了标准的绑带式机械化方程外，该 IPS 还采用了 ZUPT 技术，以防止基于 IMU 的 IPS 中典型的误差指数增长。这两种技术的结合可以减少并最大限度地降低 IMU 的固有漂移。
The main advantages of this IPS are the reduced initialization time, it is always operational, and the independence of the wearer, footwear and ground. Moreover, the system was capable of recognizing some typical first responder’s locomotion (e.g., on-the-spot turns, side/back stepping, crisscross walking and stair climbing). The authors emphasize that the results outperform the GPS ground truth, but no information is provided to assess the accuracy of the system, which is still affected by the sensor drift (especially the gyroscopes) and the magnetic perturbations. The positions of emergency responders are computed offline, so this IPS does not fulfill the requirement of the accessibility of the position information in real-time. In its current version, the IMU is too big to be mounted in the heel or insole of a boot and relies on a cable to transmit the data to the processing unit. No considerations on the system autonomy are made. 这种 IPS 系统的主要优点是初始化时间短，始终处于运行状态，而且不受穿戴者、鞋袜和地面的影响。此外，该系统还能识别急救人员的一些典型运动（如现场转弯、侧/后踏步、交叉行走和爬楼梯）。作者强调，结果优于全球定位系统的地面实况，但没有提供评估系统准确性的信息，因为系统的准确性仍然受到传感器漂移（尤其是陀螺仪）和磁扰动的影响。应急响应人员的位置是离线计算的，因此该 IPS 无法满足实时获取位置信息的要求。当前版本的 IMU 体积太大，无法安装在靴子的鞋跟或鞋垫中，只能依靠电缆将数据传输到处理单元。没有考虑系统的自主性。
The IndoorNav is an IPS designed for localization and tracking of firefighters during emergency interventions in indoor environments [82]. This system comprises three IMUs distributed on the firefighter’s body - shank, thigh and trunka RFID reader, RFID tags and a building layout. The shank IMU is designed for gait analysis, the trunk IMU to provide the orientation information and the tight IMU acceleration data, which is combined with the trunk IMU data for posture analysis. The heading estimation is provided by an Adaptive EKF and the data from the trunk IMU. IndoorNav 是一种 IPS，设计用于在室内环境中进行紧急干预时对消防员进行定位和跟踪[82]。该系统由分布在消防员身体上的三个 IMU（小腿、大腿和躯干）、RFID 阅读器、RFID 标签和建筑物布局组成。小腿 IMU 设计用于步态分析，躯干 IMU 提供方向信息和紧身 IMU 加速度数据，这些数据与躯干 IMU 数据相结合用于姿态分析。航向估计由自适应 EKF 和来自躯干 IMU 的数据提供。
Based on the data gathered from the three IMUs, the system is capable of identifying the posture and characteristic movements of firefighters-forward walking, stair climbing, stair descent forwards and stair descent backwards. The gait analysis occurs after the step detection and is based on a set of 26 fuzzy rules. The posture analysis allows the generation of alarms if a firefighter is lying down or motionless for a period of time. The RFID tags are deployed on doors and stairways by the first teams entering a building. The location coordinates of doors and stairs are stored in a database and are calculated based on the building layout. The information of the IMUs 根据从三个 IMU 收集到的数据，该系统能够识别消防员的姿势和特征动作--向前行走、爬楼梯、向前下楼梯和向后下楼梯。步态分析在步骤检测之后进行，基于一套 26 条模糊规则。如果消防员在一段时间内躺下或一动不动，姿态分析就会发出警报。射频识别（RFID）标签由第一批进入大楼的人员贴在门上和楼梯上。门和楼梯的位置坐标存储在数据库中，并根据建筑布局进行计算。IMU 的信息
and the RFID tags is fused by an EKF filter that estimates the orientation and position errors, relocates the trajectory based on the 3D coordinates of the RFID tags and corrects the heading based on a predetermined orientation when a person gets through a door. EKF 滤波器可估算方向和位置误差，根据 RFID 标签的三维坐标重新定位轨迹，并在人通过门时根据预定方向修正航向。
The reported accuracy lays above 5 meters in 90%90 \% of the measurements, which does not meet the requirements of an IPS for emergency responders. Information on how the position is displayed to firefighters and incident commander is not provided. Since the location coordinates of doors and stairs have to be calculated beforehand, this IPS depends on prior data collection. The coverage and scalability of the system are restricted by the information detail of the building layout and the RFID database. The IMU modules are connected to the processing unit through cables, decreasing the robustness, modularity and wearability of the solution and making the assemblage of the components difficult. No considerations on the system autonomy are made. 90%90 \% 报告的测量精度在 5 米以上，这不符合应急响应人员对 IPS 的要求。没有提供有关如何向消防员和事故指挥官显示位置的信息。由于必须事先计算门和楼梯的位置坐标，因此该 IPS 依赖于事先的数据收集。系统的覆盖范围和可扩展性受到建筑布局和 RFID 数据库信息细节的限制。IMU 模块通过电缆连接到处理单元，降低了解决方案的坚固性、模块化和耐磨性，并使组件组装变得困难。没有考虑系统的自主性。
The Personal Dead-Reckoning system was designed to monitor the position of emergency responders inside buildings [83]. To compute the position of the emergency responder, this IPS relies on the data of an IMU attached to the user’s boot and on three modules (position estimation, step detection and ZUPT modules). The position estimation module converts the body acceleration into the navigation reference frame, computes the linear displacement between footfalls and estimates the heading based on the data from the triad of gyroscopes. To handle with the tilt angles singularities (+-90^(@))\left( \pm 90^{\circ}\right), the position estimation module relies on Quaternion equations. The step detection module uses the angular velocities from the triad of gyroscopes to detect footfalls, for the application of the ZUPT which is used to control the drift error of the IMU. 个人惯性导航系统（Personal Dead-Reckoning system）是为监测建筑物内应急人员的位置而设计的[83]。为了计算应急响应人员的位置，该 IPS 依赖于连接在用户靴子上的 IMU 数据和三个模块（位置估计模块、脚步检测模块和 ZUPT 模块）。位置估算模块将身体加速度转换为导航参考框架，计算脚步之间的线性位移，并根据陀螺仪的三重数据估算航向。为了处理倾斜角奇异性 (+-90^(@))\left( \pm 90^{\circ}\right) ，位置估算模块依赖于四元数方程。步进检测模块使用陀螺仪三元组的角速度来检测脚步，以应用 ZUPT（用于控制 IMU 的漂移误差）。
The average error reported is below 2.1%2.1 \% of the total distance traveled, but, in extreme modes of legged locomotion - e.g., running, jumping and climbing - , the performance of the system degrades. Moreover, since the IPS does not have error control mechanisms for the orientation estimation, the heading error will grow over time. No information is provided on how the position is displayed to firefighters and incident commander and how the communication between them is done. 报告的平均误差低于总行程的 2.1%2.1 \% ，但在腿部运动的极端模式下，例如跑步、跳跃和攀爬，系统的性能会下降。此外，由于 IPS 没有方位估计误差控制机制，航向误差会随着时间的推移而增大。至于如何向消防员和事故指挥员显示位置以及他们之间如何进行通信，没有提供任何信息。
This IPS is independent of building infrastructure and prior data collection, but needs a communication infrastructure available to send the position information to the incident commander. The IMU is too big to be assembled on the emergency responder’s boot and a cable is used to connect the IMU to the processing unit, decreasing the physical robustness and the modularity of the system. The system is battery powered, but no considerations on the autonomy are made. 这种 IPS 不受建筑基础设施和事先数据收集的影响，但需要通信基础设施将位置信息发送给事故指挥官。由于 IMU 体积太大，无法安装在应急响应人员的靴子上，因此需要使用电缆将 IMU 与处理单元连接起来，从而降低了系统的物理坚固性和模块化程度。该系统由电池供电，但没有考虑自主性。
Widyawan et al. [84] proposed an IPS for the localization of firefighters and other first responders in rescue scenarios. This system was developed under the WearIT@Work project aiming a framework based on a particle filter to fuse the data from an IMU, attached to the firefighter’s boot, and building plans. The motion information is acquired from the IMU and relies on the PDR technique. Based on the building plans, a Backtracking Particle Filter (BPF) is proposed to constrain the motion of the particles and provide heading fixes. To cope Widyawan 等人[84] 提出了一种 IPS，用于在救援场景中对消防员和其他急救人员进行定位。该系统是在 WearIT@Work 项目下开发的，旨在利用基于粒子滤波器的框架，将消防员靴子上的 IMU 数据与建筑平面图融合在一起。运动信息从 IMU 获取，并依赖于 PDR 技术。根据建筑平面图，提出了一种回溯粒子滤波器（BPF）来限制粒子的运动并提供航向固定。为了应对
with limited available information on the building layout, in a simplified version of the framework, only the external building walls are considered. In this scenario, the reported mean accuracy in 2 D estimates is 1.89 meters that is lower than the PDR-only approach ( 8.04 meters mean 2D error). If more details of the building layout are added, the accuracy of the IPS increases ( 1.32 meters mean 2D error). 在建筑布局信息有限的情况下，简化版框架只考虑了建筑外墙。在这种情况下，报告的二维估算平均精度为 1.89 米，低于纯 PDR 方法（二维平均误差为 8.04 米）。如果增加更多的建筑布局细节，IPS 的精确度会提高（平均二维误差为 1.32 米）。
Woodman and Harle [85] proposed a variant of the previous system, adding the vertical information (by representing the stairways) to the buildings plans. Woodman 和 Harle [85] 提出了前一种系统的变体，在建筑平面图中增加了垂直信息（通过表示楼梯）。
With this approach, the system is capable of localizing and tracking firefighters in multi-storey buildings with a mean accuracy of 0.73 meters, in 95%95 \% of the measurements. The localization algorithm runs offline, so neither the incident commander nor the emergency responders have access to the position information in real-time. A communication infrastructure to send the position information to the incident commander was not developed/proposed. 采用这种方法，该系统能够在多层建筑中定位和跟踪消防员，在 95%95 \% 的测量中，平均精度为 0.73 米。定位算法离线运行，因此事故指挥官和应急响应人员都无法实时获取位置信息。向事故指挥官发送位置信息的通信基础设施尚未开发/提出。
This IPS is always operational, covers the entire building and is independent of the building infrastructure and prior data collection. The system performance with the typical movement patterns of the emergency responders is not addressed. Like in the previous work, the IMU is too big to be assembled on the firefighter’s boot and a cable is used to connect the IMU with a laptop. The use of cables impairs the robustness, modularity and wearability of the IPS and makes the assemblage of the components difficult. System’s autonomy is not addressed. 该 IPS 始终处于运行状态，覆盖整栋大楼，不受大楼基础设施和先前数据收集的影响。系统在应急响应人员典型运动模式下的性能问题并未涉及。与之前的工作一样，由于 IMU 体积太大，无法安装在消防员的靴子上，因此使用电缆将 IMU 与笔记本电脑连接起来。电缆的使用损害了 IPS 的坚固性、模块化和可穿戴性，并使组件的组装变得困难。系统的自主性没有得到解决。
The HeadSLAM system was developed for the localization of emergency responders and the generation of 2D environment maps, during urban search and rescue missions in unknown indoor environments [86]. Based on the particle filters and the SLAM methods, this IPS fuses the data from a head-mounted IMU and a laser range sensor to compute the emergency responder’s position and generate maps of the environment. The measurements of the laser scanner are used to estimate the orientation of the motion and the distance to obstacles (e.g., walls and furniture). The map produced by the SLAM method is similar to a building layout. HeadSLAM 系统是为在未知室内环境中执行城市搜救任务时定位应急响应人员并生成二维环境地图而开发的[86]。基于粒子滤波器和 SLAM 方法，该 IPS 融合了来自头戴式 IMU 和激光测距传感器的数据，以计算应急响应人员的位置并生成环境地图。激光扫描仪的测量值用于估算运动方向和与障碍物（如墙壁和家具）的距离。SLAM 方法生成的地图类似于建筑物布局图。
In open areas, like rooms with few obstacles, the reported accuracy is 12 meters. The accuracy of the IPS increases with the presence of obstacles ( 6 meters for corridors), but the position error is still too high for emergency responders. Like in previous works, no information is provided on how the position is displayed to firefighters and incident commander and how the communication between the firefighters and incident commander is done. 在开放区域，如障碍物较少的房间，报告的精确度为 12 米。随着障碍物的存在，IPS 的精确度会提高（走廊为 6 米），但对于应急响应人员来说，位置误差仍然过大。与之前的研究一样，没有提供关于如何向消防员和事故指挥员显示位置以及消防员和事故指挥员之间如何通信的信息。
This IPS is independent of the building infrastructure and prior data collection. The current version is too big to be used in real urban search and rescue missions. The autonomy of the system is not considered. 该 IPS 不受建筑基础设施和先前数据收集的影响。目前的版本过于庞大，无法用于实际的城市搜救任务。系统的自主性没有考虑在内。
The Computer Aided Disaster Management System (CADMS) is an emergency management system designed to support first responders in emergency scenarios [87]-[89]. The CADMS comprises an indoor positioning system, a user interface and communications facilities. The indoor positioning system integrates two software components to compute the first responder’s position. The first component processes the 计算机辅助灾难管理系统（CADMS）是一个应急管理系统，旨在为紧急情况下的第一响应者提供支持[87]-[89]。CADMS 由室内定位系统、用户界面和通信设施组成。室内定位系统集成了两个软件组件，用于计算第一响应者的位置。第一个组件处理
raw data acquired from a foot-mounted IMU and calculates the relative coordinates of the first responder. The EKF and the ZUPT methods are used, respectively, for the stabilization of the orientation estimates and the displacement error control. In the second component, the first responder’s relative coordinates are converted into global coordinates. Then, a BIM is combined with a map matching method to validate and correct, if necessary, the estimated position. The validation and correction of the first responder’s position obey to the following map matching heuristics: (1) a building wall is a natural boundary for walked paths which humans cannot pass through; (2) for a human to pass from one room to another, he/she has to use doors and the angle of passage cannot be too acute; and (3) the changes in z-coordinate only occur in stairs or ramps. These heuristics define some proprieties of the paths walked by humans. 通过脚踏式 IMU 获取原始数据，计算第一响应者的相对坐标。EKF 和 ZUPT 方法分别用于稳定方向估计值和控制位移误差。在第二部分中，第一响应者的相对坐标被转换为全局坐标。然后，将 BIM 与地图匹配方法相结合，对估计位置进行验证和修正（如有必要）。第一响应者位置的验证和修正遵循以下地图匹配启发式方法：(1) 建筑物的墙壁是步行路径的自然边界，人类无法穿过；(2) 人类要从一个房间到达另一个房间，必须使用门，并且通过的角度不能太陡；(3) Z 坐标的变化只发生在楼梯或斜坡上。这些启发法确定了人类行走路径的一些特性。
The reported mean accuracy of the system is above 1.7 meters, in 95%95 \% of the measurements. To not interfere with the activities of the rescue teams, an HMD is used to display the position and other useful information to the first responders. A MANET architecture was adopted for the communication between first responders and with the incident commander. The requirement of independence of prior data collection is not fulfilled because the BIM is site-specific and has to be created beforehand. 据报告，在 95%95 \% 的测量中，该系统的平均精度高于 1.7 米。为了不干扰救援队的活动，使用了 HMD 向急救人员显示位置和其他有用信息。第一响应者之间以及与事故指挥官之间的通信采用了城域网架构。由于 BIM 因地制宜，必须事先创建，因此无法满足事先不收集数据的要求。
In the current version, the IMU is too big to be assembled on the first responder’s boot and is connected to the processing unit by a cable. Therefore, this solution does not meet the requirements in terms of modularity, physical robustness, size, and ease of assembling. System’s autonomy is not addressed. 在当前版本中，IMU 体积太大，无法安装在急救人员的靴子上，只能通过电缆与处理装置连接。因此，这种解决方案在模块化、物理坚固性、尺寸和组装简便性方面都不符合要求。系统的自主性没有得到解决。
Rantakokko et al. [90] proposed an indoor and outdoor localization system for first responders and soldiers during emergency response and military urban operations. This IPS is based on the concept of multisensory systems and cooperative localization. Based on a Kalman filter, this system fuses the data from a foot-mounted IMU, a GPS receiver and peer-to-peer UWB ranging measurements. To overcome the cubic error drift in time of the IMU data, the ZUPT approach is applied. Unlike the majority of the proposed foot-mounted IMU approaches, the detection of the standstill step phase is achieved through a probabilistic approach [91]. To minimize the errors generated by the inertial sensors, a cooperative localization scheme based on peer-to-peer UWB ranging measurements is proposed. These high accuracy measurements are used to validate and correct, if necessary, the position estimates obtained from the IMUs readings. Based on the concept of cooperative localization, the 3D position error was reduced in the two experimental setups analyzed. Rantakokko 等人[90]提出了一种室内和室外定位系统，用于应急响应和城市军事行动中的急救人员和士兵。该 IPS 基于多感官系统和合作定位的概念。基于卡尔曼滤波器，该系统融合了来自脚踏式 IMU、GPS 接收机和点对点 UWB 测距测量的数据。为了克服 IMU 数据在时间上的立方误差漂移，采用了 ZUPT 方法。与大多数脚踏式 IMU 方法不同的是，静止步进阶段的检测是通过概率方法实现的 [91]。为尽量减少惯性传感器产生的误差，提出了一种基于点对点 UWB 测距测量的合作定位方案。这些高精度测量用于验证和修正（如有必要）从 IMU 读数中获得的位置估计值。基于合作定位的概念，在分析的两个实验装置中，三维位置误差都有所减少。
When compared with the IMU-only approach, the 3D position error was reduced from 8.2 to 3.9 meters and from 9.0 to 1.4 meters, but the tests performed are limited for a complete assessment of the IPS accuracy. How the position information is displayed to both the first responders and incident commander is not addressed. 与仅使用 IMU 的方法相比，三维位置误差从 8.2 米减小到 3.9 米，从 9.0 米减小到 1.4 米。如何向急救人员和事故指挥官显示位置信息的问题也未涉及。
This system is independent of the building infrastructure and prior data collection. In the current version, both the IMU and the UWB transceiver are too big to be assembled, respectively, on the boot and garment of the first responder. 该系统不受建筑基础设施和先前数据收集的影响。在当前版本中，IMU 和 UWB 收发器都太大，无法分别安装在第一响应者的靴子和衣服上。
Additionally, the IMU is connected to the processing unit by a cable. Therefore, this solution does not meet the requirements in terms of modularity, physical robustness, size and ease of assembling. The authors highlighted the need for efficient energy sources and pointed out the use of energy harvesting techniques to tackle energy issues, but possible implementation approaches or system’s autonomy are not considered. 此外，IMU 通过电缆与处理单元相连。因此，这种解决方案在模块化、物理坚固性、尺寸和组装简便性方面都不符合要求。作者强调了对高效能源的需求，并指出可以使用能量收集技术来解决能源问题，但没有考虑可能的实施方法或系统的自主性。
Zhang et al. [92] proposed an IPS for emergency responders based on inertial sensors attached to different body segments for urban search and rescue missions. A modified Kalman filter is used to provide a long-term stable orientation by fusing the data from the different sensors and detecting the magnetic field disturbances. To minimize the cubic error drift in time, a modified ZUPT approach is implemented on a foot-mounted IMU. The standstill phase is detected based on the gyroscopes measurements and the threshold is dynamically selected based on the velocity of the legged locomotion [93]. For the motion monitoring and selection of the ZUPT threshold, the IPS relies on the data of a chest-mounted IMU. Additionally, the system is capable of monitoring the body motion-walking and running based on IMUs mounted on each body segment. In the walking test, the accumulated error in X-Y-Z\mathrm{X}-\mathrm{Y}-\mathrm{Z} directions was 0.973m,-1.11m0.973 \mathrm{~m},-1.11 \mathrm{~m} and -0.645 m , respectively. The root mean square error (RMSE) of the orientation average over time in Roll-Pitch-Yaw directions was 0.11rad,0.05rad0.11 \mathrm{rad}, 0.05 \mathrm{rad} and 0.17 rad respectively. For the running test, the accumulated error in X-Y-Z\mathrm{X}-\mathrm{Y}-\mathrm{Z} directions were 2.35m,-1.39m2.35 \mathrm{~m},-1.39 \mathrm{~m}, and 7.73 m respectively. The orientation mean over time of RMSE in Roll - Pitch - Yaw directions were 0.35rad,0.11rad0.35 \mathrm{rad}, 0.11 \mathrm{rad} and 0.32 rad , respectively. The presentation of the position information to the first responders and incident commander is not considered. Zhang 等人[92]提出了一种基于连接在不同身体部位的惯性传感器的紧急救援人员 IPS，用于城市搜索和救援任务。该系统采用改进的卡尔曼滤波器，通过融合来自不同传感器的数据和检测磁场干扰来提供长期稳定的定位。为了最大限度地减少立方误差随时间的漂移，在脚踏式 IMU 上采用了改进的 ZUPT 方法。根据陀螺仪的测量结果检测静止阶段，并根据腿部运动的速度动态选择阈值[93]。对于运动监测和 ZUPT 阈值的选择，IPS 依赖于胸前安装的 IMU 的数据。此外，该系统还能通过安装在每个身体部位的 IMU 监测身体运动--行走和跑步。在步行测试中， X-Y-Z\mathrm{X}-\mathrm{Y}-\mathrm{Z} 方向的累积误差分别为 0.973m,-1.11m0.973 \mathrm{~m},-1.11 \mathrm{~m} 和-0.645 m。在滚动-俯仰-偏航方向上，随着时间的推移方向平均值的均方根误差（RMSE）分别为 0.11rad,0.05rad0.11 \mathrm{rad}, 0.05 \mathrm{rad} 和0.17 rad。在跑步测试中， X-Y-Z\mathrm{X}-\mathrm{Y}-\mathrm{Z} 方向的累积误差分别为 2.35m,-1.39m2.35 \mathrm{~m},-1.39 \mathrm{~m} 和 7.73 米。滚动-俯仰-偏航方向的方位均方根误差随着时间的推移分别为 0.35rad,0.11rad0.35 \mathrm{rad}, 0.11 \mathrm{rad} 和0.32 rad。未考虑向急救人员和事故指挥官提供位置信息。
This IPS is only based on IMUs mounted on different body segments, consequently, is independent of the building infrastructure and prior data collection. The IMUs used in the current implementation are connected to the processing unit with cables and are too big to be assembled on the first responder protective garment. Therefore, this solution does not meet the requirements in terms of modularity, physical robustness, size, and ease of assembling. No considerations on the autonomy of the system are made. 这种 IPS 仅以安装在不同身体部位的 IMU 为基础，因此不受建筑基础设施和事先数据收集的影响。目前实施中使用的 IMU 通过电缆连接到处理单元，而且体积太大，无法装配到急救人员的防护服上。因此，这种解决方案在模块化、物理坚固性、尺寸和组装简便性方面都不符合要求。没有考虑系统的自主性。
Hari et al. [94] proposed an IPS to support first responders in disaster assistance operations, comprising two footmounted IMUs, a helmet-mounted camera, a processing platform strapped around the waist and a UWB transceiver for accurate peer-to-peer UWB ranging measurements. The IMUs are assembled into the heels of the shoe and the inertial navigation modules are based on the OpenShoe project [95], [96]. The data collected from the IMUs is processed with a nonlinear inequality filtering approach that limits the maximum foot-to-foot distance to 1 meter [97]. The step-to-step displacement is calculated based on the ZUPT approach. Based on the concept of cooperative localization, inter-agent ranging measurements are used to refine the position estimate. The implementation details of centralized cooperative localization algorithms based on dual foot-mounted IMUs and inter-agent ranging are described in [98]. Hari等人[94]提出了一种IPS，用于支持灾难援助行动中的急救人员，它由两个安装在脚上的IMU、一个安装在头盔上的摄像头、一个绑在腰部的处理平台和一个用于精确点对点UWB测距测量的UWB收发器组成。IMU 安装在鞋跟处，惯性导航模块基于 OpenShoe 项目[95]、[96]。从 IMUs 收集到的数据采用非线性不等式滤波方法进行处理，将脚到脚的最大距离限制在 1 米[97]。步间位移根据 ZUPT 方法计算。根据合作定位的概念，利用机构间的测距测量来完善位置估计。基于双脚踏式 IMU 和代理间测距的集中式合作定位算法的实施细节在 [98] 中进行了描述。
The reported average position accuracy is about 2 meters. The authors refer that the helmet-mounted camera can be used for SLAM approaches, but in the current solution it is only used to record the intervention. The position information display to both the first responders and incident commander is not considered. 报告的平均定位精度约为 2 米。作者提到，头盔上的摄像头可用于 SLAM 方法，但在当前的解决方案中，它仅用于记录干预过程。没有考虑向急救人员和事故指挥官显示位置信息。
The pre-existing communication network is used for communication with the incident commander. In the case of grid breakdown, the IPS will not be capable of communicating. Each IMU communicates with the processing platform via Bluetooth. These hardware features fulfill the IPS requirements of modularity, size, physical robustness, and ease of assembling. No information about the processing platform is provided. The IMUs are battery powered and have an autonomy of 1.5 hours, which does not fulfill the requirements specified in Section III-E. 原有的通信网络用于与事故指挥官通信。在电网瘫痪的情况下，IPS 将无法进行通信。每个 IMU 都通过蓝牙与处理平台通信。这些硬件功能满足了 IPS 对模块化、尺寸、物理坚固性和组装简便性的要求。没有提供有关处理平台的信息。IMU 由电池供电，续航时间为 1.5 小时，不符合第 III-E 节中规定的要求。
The REFIRE solution aims to provide localization and tracking support to first responders in emergency response scenarios [99]. The system comprises Mobile Terminals (MTs) carried by the first responders, low-cost Pre-Installed Location Devices (PILDs) embedded into the existing safety devices (e.g., emergency lights and smoke detectors) and a Command and Control Center (CCC) for the mission coordination. Each MT integrates a RFID reader, an IMU mounted at the pelvis level and a processing unit. The computed first responder’s position is sent to the CCC through 2G//3G//4G2 \mathrm{G} / 3 \mathrm{G} / 4 \mathrm{G} wireless networks. The PILDs are passive RFID tags preinstalled on the building and, in each one of them, the fixed location data is stored in the memory bank based on the WGS-84 standard. Other key information can be stored, e.g., presence of dangerous materials, room type and indications to the nearest exit. REFIRE 解决方案旨在为应急响应场景中的急救人员提供定位和跟踪支持[99]。该系统由急救人员携带的移动终端（MT）、嵌入现有安全设备（如应急灯和烟雾探测器）的低成本预装定位设备（PILD）以及用于任务协调的指挥控制中心（CCC）组成。每个 MT 都集成了一个 RFID 阅读器、一个安装在骨盆位置的 IMU 和一个处理单元。计算出的第一响应者位置通过 2G//3G//4G2 \mathrm{G} / 3 \mathrm{G} / 4 \mathrm{G} 无线网络发送到 CCC。PILD 是预先安装在建筑物上的无源 RFID 标签，每个标签中的固定位置数据都存储在基于 WGS-84 标准的存储库中。还可存储其他关键信息，如危险品的存在、房间类型和最近出口的指示。
To achieve both resilience and a meter-level accuracy the IPS is dimensioned [27] to a maximum range and angle of each RFID tag of 3 meters and 120^(@)120^{\circ}, respectively. The Rescuer Localization Algorithm (RLA) [13] exploits the IMU and RFID data to estimate the position of a first responder. The PDR approach is used to compute the location and heading. For the heading estimation, an EKF is used to fuse the data from the triad of accelerometers, gyroscopes and magnetometers. The heading estimate is combined with the accelerometer readings to compute the vertical accelerations, which are associated with the rise and fall of the pelvis during a gait cycle and are used to compute the step length. An adaptive time windows technique is used to detect the initial contact of each step. The position estimated based on the IMU data is then refined with the RFID data, if available. 为了实现弹性和米级精度，IPS 的尺寸[27]为每个 RFID 标签的最大范围和角度分别为 3 米和 120^(@)120^{\circ} 。救援人员定位算法（RLA）[13] 利用 IMU 和 RFID 数据来估计第一响应者的位置。PDR 方法用于计算位置和航向。在估计航向时，使用 EKF 融合加速度计、陀螺仪和磁力计三者的数据。航向估计值与加速度计读数相结合来计算垂直加速度，垂直加速度与步态周期中骨盆的上升和下降有关，用于计算步长。自适应时间窗技术用于检测每一步的初始接触。然后，根据 IMU 数据估算出的位置再与 RFID 数据（如果有的话）进行完善。
This IPS does not perform correction on the attitude of the first responder. The best-reported accuracy of this IPS is 4 meters, which does not fulfill the requirements specified in Section III-A. The position information is accessible to both the incident commander and the first responder through the CCC and the MT, respectively. In the current version, the position of the first responders is computed offline. 该 IPS 不对第一反应者的姿态进行校正。据报告，该 IPS 的最佳精度为 4 米，不符合第 III-A 节规定的要求。事故指挥官和第一响应者可分别通过 CCC 和 MT 获取位置信息。在当前版本中，第一响应者的位置是离线计算的。
The system depends on the building infrastructure, as the RFID tags have to be preinstalled and their position must be calculated and stored manually during deployment. Additionally, a parameter-specific of the first responder’s body 该系统依赖于楼宇基础设施，因为必须预先安装 RFID 标签，并且在部署过程中必须手动计算和存储标签的位置。此外，第一响应者身体的特定参数
has to be calculated beforehand for the step length computation, making the IPS dependent on prior data collection. The coverage is also defined at the deployment stage and no changes are allowed during emergency operations. The system’s autonomy depends on both the RFID reader and the MT autonomies, which are individually battery powered. 必须事先计算步长，因此 IPS 依赖于事先收集的数据。覆盖范围也是在部署阶段确定的，在紧急行动期间不允许更改。系统的自主性取决于 RFID 阅读器和 MT 的自主性，它们都由电池单独供电。
Rantakokko et al. [19] proposed an IPS to support firefighters during smoke diving operations comprising two footmounted IMUs and one knee-mounted IMU. The data from the IMUs is fused to provide a robust position estimation of firefighters in typical movements: walk (upright or hunched), “knee-dragging motion” or crawling. The knee-mounted IMU applies the ZUTP during the crawling motion. In both IMU configurations (foot- and knee-mounted), the zero-velocity detection is based on a threshold value. This threshold is defined by the RMS value of the angular velocity during a period of five samples. For sensor fusion, the constrained Kalman filter proposed by Skog et al. [97] was adopted. This sensor fusion approach defines an upper bound on maximum spatial separation to 1 meter. Rantakokko 等人[19] 提出了一种 IPS，用于在烟雾潜水行动中为消防员提供支持，该系统由两个脚部安装的 IMU 和一个膝部安装的 IMU 组成。通过融合 IMU 的数据，可对消防员的典型运动进行稳健的位置估计：步行（直立或驼背）、"膝拖运动 "或爬行。膝部安装的 IMU 在爬行运动中应用 ZUTP。在两种 IMU 配置（脚踏式和膝装式）中，零速度检测都基于一个阈值。该阈值由五个采样周期内的角速度均方根值定义。在传感器融合方面，采用了 Skog 等人提出的约束卡尔曼滤波器[97]。这种传感器融合方法将最大空间间隔的上限定义为 1 米。
This system demonstrates that with foot- and knee-mounted IMUs in both legs it is feasible to compute the position of firefighters for all movements with the required accuracy. The authors do not refer how the position information is displayed to both the firefighters and the incident commander. 该系统表明，通过在双腿的脚部和膝部安装 IMU，可以计算出消防员所有动作的位置，并达到所需的精度。作者没有提及如何向消防员和事故指挥官显示位置信息。
This IPS is independent of the building structure and prior data collection. In the current version, the IMUs are connected to the processing unit with cables and are too big to be assembled on the protective garment. Therefore, this solution does not meet the requirements in terms of modularity, physical robustness, size, and ease of assembling. System’s autonomy is not addressed. 这种 IPS 不受建筑结构和事先数据收集的影响。在当前版本中，IMU 是用电缆连接到处理单元的，而且体积太大，无法装配到保护服上。因此，该解决方案在模块化、物理坚固性、尺寸和组装简便性方面都不符合要求。系统的自主性没有得到解决。
The Tactical lOcatoR (TOR) system is an IPS designed for the localization of firefighters during search operations [100], integrating a dual foot-mounted IMU and a UWB radio tag for inter-agent ranging measurements. The IMUs (OpenShoe IMUs [95], [96]), the fusing algorithm of the dual footmounted IMU [97] and the centralized cooperative localization algorithm [98] are similar to the proposed by Hari et al. [94]. The main difference is the UWB-based inter-agent ranging measurements. This IPS relies on the preinstalled Ubisense system to obtain the synthetic ranging measurements from the positions estimates [2]. The cooperative localization algorithm is centralized, but the sensor fusion algorithm of the dual foot-mounted IMUs is decentralized, allowing the system to operate continuously during significant time periods without the inter-agent ranging measurements and contact with the central node. 战术定位系统（TOR）是专为搜索行动中的消防员定位而设计的 IPS [100]，集成了一个双脚踏式 IMU 和一个 UWB 无线电标签，用于机构间测距测量。该系统的 IMU（OpenShoe IMU [95], [96]）、双脚踏式 IMU 的融合算法 [97] 和集中式合作定位算法 [98] 与 Hari 等人提出的系统类似 [94]。主要区别在于基于 UWB 的代理间测距测量。这种 IPS 依靠预装的 Ubisense 系统从位置估计值中获取合成测距数据[2]。合作定位算法是集中式的，但双脚踏式 IMU 的传感器融合算法是分散式的，这使得系统可以在相当长的时间内连续运行，而无需进行代理间测距测量，也无需与中央节点联系。
The IPS was evaluated in walking and crawling movements and the reported accuracy is 1 and 3 meters, respectively. The processing platform is a smartphone that runs the fusion algorithm in real-time and calculates a new position estimate every second. The position and heading information are transmitted to the command and control system via IEEE 802.11 WLAN radio links. The presentation of position information to first responders and incident commander is not mentioned. Each foot-mounted IMU is connected to a smartphone through cables and is battery powered (autonomy 1.5 hours). These IPS 在行走和爬行运动中进行了评估，报告的精度分别为 1 米和 3 米。处理平台是一部智能手机，实时运行融合算法，每秒计算一次新的位置估计值。位置和航向信息通过 IEEE 802.11 WLAN 无线链路传输到指挥和控制系统。向急救人员和事故指挥员提供位置信息的方式没有提及。每个脚踏式 IMU 都通过电缆与智能手机连接，并由电池供电（续航时间为 1.5 小时）。这些
characteristics do not fulfill the IPS requirements in terms of modularity, physical robustness, ease of assembling and the autonomy. The smartphone and the Ubisense radio tag are also battery powered. The autonomy of the system is not considered. 在模块化、物理坚固性、组装简便性和自主性方面，这些特点都不符合 IPS 的要求。智能手机和 Ubisense 无线电标签也由电池供电。系统的自主性没有考虑在内。
Summary of IMU-based Systems: Unlike the radio signalbased systems, an IMU-based IPS has zero-radiation signature which makes it “invisible” to sensors in hostile environments and immune to interference or jamming. Therefore, the security of the position information is guaranteed. The performance of these IPSs is not affected by the quantity and type of materials of the walls or by the phenomena that affect the propagation of RF signals. Moreover, the system is capable of providing position estimations in all indoor environments. 基于 IMU 的系统概述：与基于无线电信号的系统不同，基于 IMU 的 IPS 具有零辐射特征，这使其对恶劣环境中的传感器 "不可见"，并且不受干扰或干扰。因此，位置信息的安全性得到了保证。这些 IPS 的性能不受墙壁材料数量和类型的影响，也不受影响射频信号传播的现象的影响。此外，该系统还能在所有室内环境中提供位置估计。
The cumulative error associated with these IPSs is a major issue, still to be solved. This error is associated with the IMUs drift and grows exponentially due to twofold integration required to obtain the displacement from the acceleration measurements, for instance. Therefore, without a periodic position correction, the error of the position estimation will be so high that this information will be useless to the emergency responders and the incident commander. 与这些 IPS 相关的累积误差是一个有待解决的主要问题。这种误差与 IMU 的漂移有关，并且由于从加速度测量中获取位移所需的两倍积分等原因而呈指数增长。因此，如果不进行定期位置校正，位置估算的误差将非常大，以至于这些信息对应急响应人员和事故指挥官毫无用处。
Many approaches have been proposed to minimize the impact of the IMUs drift, but for long term operations this phenomenon still has a negative impact on the IPS performance. 人们提出了许多方法来尽量减少 IMUs 漂移的影响，但对于长期运行来说，这种现象仍然会对 IPS 性能产生负面影响。
The type of motion is another factor affecting the accuracy of the position estimate. Unlike regular PDR applications, where the wearer’s movements are walking, running, climbing and descending stairs, the typical movements of emergency responders during their missions - walk (upright or hunched), “knee-dragging motion” and crawling - are much more challenging and need further research. Some preliminary studies have been conducted on these mission-specific movements, but none of the existing IPSs is capable of calculating the emergency responder’s position regardless the movement involved. The lack of a communication infrastructure to send the computed position to the incident commander is a further drawback of this approach. 运动类型是影响位置估计准确性的另一个因素。在普通的 PDR 应用中，佩戴者的运动包括行走、跑步、爬楼梯和下楼梯，而应急响应人员在执行任务时的典型运动--行走（直立或驼背）、"拖膝运动 "和爬行--则更具挑战性，需要进一步研究。针对这些特定任务的动作已经进行了一些初步研究，但现有的 IPS 都无法计算应急响应人员的位置，无论涉及何种动作。这种方法的另一个缺点是缺乏通信基础设施，无法将计算出的位置发送给事故指挥官。
C. Hybrid Systems C.混合系统
As highlighted in the two previous sections, the selection between Radio Signal and IMU-based technologies for the localization process has some implications on the IPS performance. On the one hand, the low availability of RF signals in indoor environments is the major limitation of the radio signal-based IPSs. On the other hand, in IMU-based IPSs, the position of the emergency responder can be continuously calculated, but the inherent drift of the sensors that compose an IMU degrades the accuracy of the position estimate over the time. To overcome these limitations, several researchers proposed IPSs that combine both approaches in a single solution. In a hybrid IPS, the two subsystems (radio signal- and IMU-based) calculate the first responder’s position independently and then, a data fusion algorithm merges both data to provide a single position estimate. In this section, the 正如前面两节所强调的，在定位过程中选择无线电信号技术还是基于 IMU 的技术对 IPS 性能有一定影响。一方面，射频信号在室内环境中的可用性较低是基于无线电信号的 IPS 的主要局限性。另一方面，在基于 IMU 的 IPS 中，可以连续计算应急响应者的位置，但组成 IMU 的传感器的固有漂移会随着时间的推移降低位置估计的准确性。为了克服这些限制，一些研究人员提出了将两种方法结合在一个解决方案中的 IPS。在混合式 IPS 中，两个子系统（基于无线电信号和 IMU）分别计算第一响应者的位置，然后通过数据融合算法将两个数据合并，以提供单一的位置估计值。在本节中
IPSs for emergency responders that rely on both technological approaches are introduced and discussed. 介绍并讨论了依靠这两种技术方法为应急响应人员提供的 IPS。
The Relate Trails is an IPS for emergency responders during search and rescue missions [21], [101], [102]. This IPS combines the ultrasound technology with a foot-mounted IMU for the localization of emergency responders in unknown environments. Relate Trails 是一种用于搜救任务中应急响应人员的 IPS [21]、[101]、[102]。该 IPS 将超声波技术与脚踏式 IMU 相结合，用于在未知环境中确定应急响应人员的位置。
The ultrasound beacons are deployed, acting as waypoints. These beacons create a “breadcrumb” trail that provides relative positioning, corrects the PDR estimated trajectory and gives assisted navigation to the emergency responders. The foot-mounted IMU relies on the ZUPT and a Kalman filter for continuous position estimation and sensor drift control. The standstill phase is detected by applying a threshold to the product of the acceleration norms by the rate of turn. Since the navigation functionality depends only on the wearer’s relative position to the nearest waypoint, it is not affected by the error of the position estimated based on the PDR approach. On the other hand, the use of an IMU for the localization allows calculating the emergency responder’s position even when no waypoints are available (e.g., waypoints destroyed by the fire). 部署超声波信标，作为航点。这些信标形成的 "面包屑 "轨迹可提供相对定位，修正 PDR 估计轨迹，并为应急响应人员提供辅助导航。脚踏式 IMU 依靠 ZUPT 和卡尔曼滤波器进行连续位置估算和传感器漂移控制。通过对加速度与转弯率的乘积应用阈值来检测静止阶段。由于导航功能只取决于佩戴者与最近航点的相对位置，因此不会受到基于 PDR 方法估计的位置误差的影响。另一方面，使用 IMU 进行定位，即使在没有航点（例如航点被大火烧毁）的情况下，也能计算出紧急救援人员的位置。
The error in the position estimate can reach up to eight percent of the total distance traveled and the error in the heading estimate is not provided. This position error is too high for emergency responders’ mission scenarios, however, since the ultrasound signals do not penetrate walls, the system is also capable of providing a room-level localization. The relative position and the navigation support are displayed to the emergency responder by an HMD. The SLAM technique is used to improve the accuracy of the position of the ultrasound beacons. This IPS is independent from the building infrastructure and prior data collection and its coverage is limited by the number of ultrasound beacons carried by the emergency responders. 位置估计误差可达总行程的 8%，而且不提供航向估计误差。这种位置误差对于应急响应人员的任务场景来说太高了，不过，由于超声波信号不会穿透墙壁，该系统还能提供房间级定位。相对位置和导航支持通过 HMD 显示给应急响应人员。SLAM 技术用于提高超声信标位置的准确性。该 IPS 独立于建筑基础设施和先前的数据收集，其覆盖范围受到应急响应人员携带的超声信标数量的限制。
Despite the challenging scenarios addressed by this IPS, no considerations are made on the physical robustness of the ultrasound beacons. Moreover, the foot-mounted IMU is connected to a laptop through a cable. The use of cables does not fulfill the IPS requirements in terms of modularity, physical robustness and ease of assembling. The power source and autonomy of the system are not described. 尽管该 IPS 解决的是具有挑战性的场景，但并未考虑超声信标的物理坚固性。此外，脚踏式 IMU 通过电缆与笔记本电脑连接。电缆的使用不符合 IPS 在模块化、物理坚固性和组装简便性方面的要求。系统的电源和自主性未作说明。
The Precision Personal Locator (PPL) is an IPS developed for locating and tracking first responders and other personnel in search and rescue missions [25], [103]-[107]. The system comprises a base station, fixed transceivers mounted on trucks or ladders and a portable unit carried by the first responders. 精确个人定位器（PPL）是一种 IPS，用于定位和跟踪搜救任务中的急救人员和其他人员[25]、[103]-[107]。该系统由一个基站、安装在卡车或梯子上的固定收发器和急救人员携带的便携式装置组成。
Portable units contain an IMU and a pulsed radio transmitter to reduce the multipath error. For the localization process, both the IMU and the radio transceiver are independently used to provide an initial position estimate, which will be later fused at the base station for a final position estimate. The IMU continuously estimates the first responder’s position, which is transmitted to the base station. At the same time, the portable unit’s radio transceiver broadcasts a pulsed radio signal that will be received by the fixed transceivers. A Bayesian fusion algorithm is used to calculate the first responder’s position based on, at least, three TDoA ( sigma\sigma ART algorithm [105]) or ToA (TART algorithm [104]) measurements received from fixed transceivers. A multi-carrier wideband signal is used to 便携式设备包含一个 IMU 和一个脉冲无线电发射器，以减少多径误差。在定位过程中，IMU 和无线电收发器各自独立使用，以提供初始位置估计值，然后在基站进行融合，得出最终位置估计值。IMU 不断估算第一响应者的位置，并将其传送到基站。与此同时，便携式设备的无线电收发器会广播脉冲无线电信号，固定收发器也会接收到该信号。贝叶斯融合算法用于根据从固定收发器接收到的至少三个 TDoA（ sigma\sigma ART 算法 [105]）或 ToA（TART 算法 [104]）测量值计算第一响应者的位置。多载波宽带信号用于
measure the distance between the portable unit and a fixed transceiver. The signal structure is similar to the signals modulated with the OFDM approach. Based on a mathematical technique known as synthetic aperture imaging, the base station fuses the position information obtained from the broadcasted IMU’s position estimate and the transmitted ranging signals of the portable unit. 测量便携式设备与固定收发器之间的距离。信号结构类似于用 OFDM 方法调制的信号。根据一种称为合成孔径成像的数学技术，基站将从广播的 IMU 位置估计值和便携式设备发射的测距信号中获得的位置信息进行融合。
In residential buildings, this IPS demonstrated the ability to track multiple first responders with sub-meter accuracy [104]. However, the system was tested only on wheeled platforms [103] and requires a huge number of fixed transceivers deployed outside the building [105]. A GUI displays to the incident commander the first responders’ position, as well as, the physiological and environmental data. How the information is displayed to the first responders is not provided. The coverage is limited by the penetration of the RF signals in the building. The localization algorithm is implemented in a Field-Programmable Gate Array (FPGA) as a coprocessor, reducing the device form-factor and power consumption and increasing the device performance [107]. The IPS is battery powered and has an autonomy up to 24 hours. 在住宅楼宇中，这种 IPS 展示了以亚米级精度跟踪多名急救人员的能力[104]。不过，该系统仅在轮式平台上进行了测试 [103]，而且需要在建筑物外部署大量固定收发器 [105]。图形用户界面向事故指挥官显示第一反应人员的位置以及生理和环境数据。至于如何向急救人员显示这些信息，则没有提供。覆盖范围受到建筑物内射频信号穿透力的限制。定位算法作为协处理器在现场可编程门阵列（FPGA）中实现，从而降低了设备外形尺寸和功耗，提高了设备性能[107]。IPS 由电池供电，自主运行时间长达 24 小时。
The Virtual Lifeline is a multimodal indoor positioning and navigation system to support emergency responders during search and rescue missions [108]. This IPS comprises a footmounted IMU, Relate ultrasound nodes and RF nodes. 虚拟生命线是一种多模态室内定位和导航系统，可在搜救任务中为应急响应人员提供支持[108]。该 IPS 由脚踏式 IMU、Relate 超声波节点和射频节点组成。
The foot-mounted IMU continuously estimates the wearer’s position based on the PDR approach and the ZUPT. The standstill phase of the ZUPT is detected by applying a threshold to the product of the norms of acceleration by the rate of turn. To control the inherent drift of the IMU-based systems, the estimated position from the PDR system is periodically realigned using both types of nodes. The ultrasound nodes are used for short range high accuracy position and heading realignments. The RF nodes have a lower position accuracy, but a longer range, therefore more nodes may be detected at one location. These nodes only provide position corrections. Both nodes are deployed in a dynamic ad hoc manner, as the emergency responder enters a building. To cope with RF propagation phenomena - reflection, diffraction and fading - the RF beacons transmit at different transmission powers ( -22,-9-22,-9, and 0 dBm ). The position of each node is assigned based on the emergency responder’s current position, which is estimated based on the PDR system. A particle filter is used to fuse the information collected from the PDR, the RF transceiver (RSSI) and the ultrasound sensor. A backtracking particle filter was also implemented to refine the state estimates based on the particle trajectory historic and on removing the invalid particle trajectories. 脚踏式 IMU 根据 PDR 方法和 ZUPT 不断估算佩戴者的位置。ZUPT 的静止阶段是通过对加速度标准与转弯率的乘积应用阈值来检测的。为了控制基于 IMU 的系统的固有漂移，PDR 系统的估计位置会定期使用这两种类型的节点进行重新调整。超声波节点用于短距离高精度位置和航向调整。射频节点的位置精度较低，但射程较远，因此可以在一个位置检测到更多节点。这些节点只提供位置校正。这两种节点都是在应急响应人员进入建筑物时以动态临时方式部署的。为了应对射频传播现象（反射、衍射和衰减），射频信标以不同的发射功率（ -22,-9-22,-9 和 0 dBm）进行发射。每个节点的位置是根据应急响应者的当前位置分配的，而应急响应者的当前位置是根据 PDR 系统估算的。粒子滤波器用于融合从 PDR、射频收发器（RSSI）和超声波传感器收集到的信息。此外，还采用了回溯粒子滤波器，以根据粒子轨迹历史记录和去除无效粒子轨迹来完善状态估计。
The reported accuracy when the position of the waypoints is assigned by the PDR system is 3.88 and 8.66 meters in normal walking mode and extended search mode, respectively. If the exact coordinates of the nodes are known (nodes acting as landmarks), the accuracy of the IPS is 2 meters. Both situations are inadequate for emergency responders, as the position error is too high. No information is provided how the position is displayed to the incident commander and the emergency responders. The coverage of the system is limited by the number of nodes carried by the emergency responders. 在正常行走模式和扩展搜索模式下，由 PDR 系统指定航点位置时报告的精度分别为 3.88 米和 8.66 米。如果知道节点的确切坐标（节点充当地标），IPS 的精确度为 2 米。由于位置误差过大，这两种情况都不能满足紧急救援人员的需要。没有提供如何向事件指挥员和应急人员显示位置的信息。该系统的覆盖范围受到应急人员携带的节点数量的限制。
In the current version, the nodes are implemented on different hardware platforms, which increases the size, weight, and deployment effort of the proposed solution. Moreover, the foot-mounted IMU is connected to a laptop through a cable, which does not fulfill the IPS requirements in terms of modularity, physical robustness, and ease of assembling. No considerations about the power source and autonomy of the system are made. 在当前版本中，节点是在不同的硬件平台上实现的，这就增加了拟议解决方案的体积、重量和部署工作量。此外，脚踏式 IMU 通过电缆连接到笔记本电脑上，在模块化、物理坚固性和组装简便性方面不符合 IPS 的要求。没有考虑系统的电源和自主性。
The Geospatial Location Accountability and Navigation System for Emergency Responders (GLANSER) is an indoor and outdoor localization system to support emergency responders, firefighters in particular, during all mission types [25], [109]. This IPS comprises three parts: the Geospatial Locator Unit (GLU), the Anchor Panel Unit (APU) and the Commander Display Unit (CDU). 应急响应人员地理空间定位问责和导航系统（GLANSER）是一个室内外定位系统，用于在所有任务类型中支持应急响应人员，特别是消防员[25]，[109]。该 IPS 由三部分组成：地理空间定位单元（GLU）、锚定面板单元（APU）和指挥官显示单元（CDU）。
The GLU is responsible for the communication with the incident commander, navigation, and localization of the firefighters. The APUs are mounted on the emergency vehicles and provide support functions to the GLUs (e.g., charging, geostationary referencing, and UWB ranging measurements). The CDU is responsible for the graphical interface with the incident commander and by the graphical construction of the building. Each firefighter has a GLU attached to the breathing gear that combines a military IMU, a GPS receiver, Doppler radars to correct velocity, a pressure sensor to measure changes in altitude, and a radio transceiver - with UWB technology - to measure the range to APUs and other nearby GLUs [110]. GLU 负责与事故指挥官通信、导航和消防员定位。APU 安装在应急车辆上，为 GLU 提供支持功能（如充电、地球静止参照和 UWB 测距）。CDU 负责与事故指挥官的图形界面以及建筑物的图形构造。每个消防员的呼吸装备上都有一个 GLU，该 GLU 结合了军用 IMU、GPS 接收器、校正速度的多普勒雷达、测量高度变化的压力传感器以及采用 UWB 技术的无线电收发器，用于测量 APU 和附近其他 GLU 的距离[110]。
The synthetic aperture technique is applied to increase the accuracy of the UWB ranging measurements [111]. An EKF is used to fuse all the data acquired from the sensors embedded on the GLU. The data fusion algorithm for the localization is implemented in an FPGA to improve the system’s performance. The central module of the GLU is called GLANSER Embedded Processor (GEP). 合成孔径技术用于提高 UWB 测距测量的精度[111]。EKF 用于融合从嵌入在 GLU 上的传感器获取的所有数据。用于定位的数据融合算法由 FPGA 实现，以提高系统性能。GLU 的中心模块称为 GLANSER 嵌入式处理器（GEP）。
In the current version, the reported position accuracy is 3 meters. The UWB ranging measurements were not used in the tests, only the information acquired from the IMU and other sensors. The firefighters’ position is displayed, in realtime, to the incident commander at the CDU, but not to the firefighters. For the communication, each APU implements a hybrid mesh network, connecting up to eleven GLUs with the CDU. More firefighters - up to 500 - can be supported if more APUs are deployed. The weight of the current version complies with the defined requirements (0.45kg)(0.45 \mathrm{~kg}), but the GLU, mainly due to battery, is too bulky and heavy, interfering with the firefighters’ activities. To overcome this limitation, a thin and flexible battery that could slip into the lining of a firefighter’s jacket [25] is being developed. This IPS has an autonomy up to four hours. The main drawback of the system is the high cost. 在当前版本中，报告的定位精度为 3 米。测试中没有使用 UWB 测距测量，只使用了从 IMU 和其他传感器获得的信息。消防员的位置会实时显示给 CDU 的事故指挥官，但不会显示给消防员。在通信方面，每个 APU 都采用了混合网状网络，最多可将 11 个 GLU 与 CDU 连接起来。如果部署更多的 APU，则可支持更多的消防员（最多 500 人）。当前版本的重量符合 (0.45kg)(0.45 \mathrm{~kg}) 规定的要求，但 GLU（主要是电池）过于笨重，影响了消防员的活动。为了克服这一限制，目前正在开发一种轻薄灵活的电池，可以放入消防员外套的内衬中 [25]。这种 IPS 的自主运行时间长达四小时。该系统的主要缺点是成本较高。
The Wearable Advanced Sensor Platform (WASP) is a system developed to support firefighters during fire operations [112]. The system comprises a flame-resistant Tshirt for the acquisition of physiological parameters (e.g., heart rate, heart rate variability, respiration rate, activity levels and posture), a waist belt device from TRX NEON Systems for the indoor localization [113], a smartwatch to provide feedback to 可穿戴先进传感器平台（WASP）是为消防员在灭火行动中提供支持而开发的系统[112]。该系统包括一件用于采集生理参数（如心率、心率变异性、呼吸频率、活动水平和姿势）的阻燃 Tshirt、一个用于室内定位的 TRX NEON 系统腰带设备[113]、一个用于向消防员提供反馈的智能手表[114]、一个用于向消防员提供信息的可穿戴式先进传感器平台[115]和一个用于向消防员提供信息的可穿戴式先进传感器平台[116]。
the firefighter and a command station to process and display location and other parameters. 消防员和指挥站处理并显示位置和其他参数。
The NEON Personnel Location System is the unit responsible for calculating the firefighter’s position. This device fuses the data from inertial sensors, pressure sensor and ranging sensor, to roughly estimate the firefighter’s position and, refines that estimate based on waypoints realignments, historical track and, inferred and known map data. The waypoints are based on the Near Field Communication (NFC) technology. NEON 人员定位系统是负责计算消防员位置的装置。该装置融合了惯性传感器、压力传感器和测距传感器的数据，可以大致估算出消防员的位置，并根据航点重新定位、历史轨迹以及推断和已知地图数据对估算结果进行改进。航点基于近距离无线通信（NFC）技术。
No details about the data fusion algorithm are provided. In fact, little information is available about the performance and specifications of the system, e.g., the manufacturer does not mention the localization accuracy. A GUI and an Android Smartphone are used to display the firefighters’ position and other parameters to the incident commander and the firefighters, respectively. For the corrections of the position estimate, this IPS relies on building layouts, so it depends on prior data collection. The coverage can be expanded by deploying portable multi-sensor anchor nodes as firefighters enter a building. No information about the size, weight, physical robustness, power type and autonomy of the wearable units are given. The wearable sensors are embedded on an adjustable strap that can be unbuckled for easier donning and doffing. 没有提供数据融合算法的详细信息。事实上，关于系统性能和规格的信息也很少，例如，制造商没有提及定位精度。系统使用图形用户界面和安卓智能手机分别向事故指挥官和消防员显示消防员的位置和其他参数。对于位置估计的校正，该 IPS 依赖于建筑物布局，因此依赖于先前的数据收集。在消防员进入建筑物时部署便携式多传感器锚节点，可以扩大覆盖范围。关于可穿戴设备的尺寸、重量、物理坚固性、电源类型和自主性，我们没有提供任何信息。可穿戴传感器嵌在一条可调节的带子上，带子可以解开，方便穿脱。
Simon et al. [114] proposed an IPS to support emergency responders, in particular firefighters, during rescue operations in catastrophic scenarios. This IPS comprises a wireless footmounted IMU, a handheld device and preinstalled low-power wake-up landmarks integrated into smoke detectors throughout the building. Simon 等人[114]提出了一种 IPS，用于在灾难性场景的救援行动中为应急响应人员（尤其是消防员）提供支持。这种 IPS 包括一个无线脚踏式 IMU、一个手持设备和预装在整个建筑物烟雾探测器中的低功耗唤醒地标。
The firefighter’s position is calculated by two independent subsystems: the PDR and the landmark subsystems. The PDR subsystem is based on a micro-IMU especially developed for applications where the size, weight, and power consumption are key requirements [115]. So, these design goals are achieved by sending the raw data wirelessly and post-processing in a handheld device. By combining the ZUPT algorithm with a Kalman filter, the raw data from the IMU is fused to track the firefighter. The standstill phase is detected when the norm value of the gyroscopes outputs is smaller than a predefined threshold. 消防员的位置由两个独立的子系统计算：PDR 子系统和地标子系统。PDR 子系统基于一个微型 IMU，该系统是专为对尺寸、重量和功耗有严格要求的应用而开发的 [115]。因此，通过无线方式发送原始数据并在手持设备中进行后处理，可以实现这些设计目标。通过将 ZUPT 算法与卡尔曼滤波器相结合，融合 IMU 的原始数据来跟踪消防员。当陀螺仪输出的标准值小于预定义的阈值时，就会检测到静止阶段。
The landmark subsystem is based on the RSSI measured from the message exchanges with the landmarks. In this subsystem, the handheld device broadcasts a wake-up message and measures the RSSI of each reply message received from landmarks that woke-up. Based on the RSSI, the handheld device runs optimized localization algorithms - the Gradient Descent Method and the Gauss-Newton algorithm - to determine the current position of the handheld device. To calculate the firefighter’s position the device has to stand still for a while and the landmarks’ position must be known. The same method can also be used to compute the landmarks position. In this case, the position of the firefighter has to be known. For both subsystems, the localization algorithms run in the handheld device. 地标子系统基于与地标交换信息时测量到的 RSSI。在该子系统中，手持设备广播唤醒信息，并测量从唤醒地标收到的每个回复信息的 RSSI。根据 RSSI，手持设备运行优化的定位算法--梯度下降法和高斯-牛顿算法--来确定手持设备的当前位置。要计算消防员的位置，设备必须静止一段时间，而且必须知道地标位置。同样的方法也可用于计算地标位置。在这种情况下，必须知道消防员的位置。这两个子系统的定位算法都在手持设备中运行。
In the current version, the IMU data and the RSSI distance data are not fused. The mean error reported is 99 cm , but, the user has to stand still during the exchange of messages with the landmarks, which is unreliable during an emergency 在当前版本中，IMU 数据和 RSSI 距离数据没有融合。报告的平均误差为 99 厘米，但在与地标交换信息时，用户必须站立不动，这在紧急情况下是不可靠的。
response mission. The firefighter’s position is displayed at the handheld device. How this information is sent and displayed to the incident commander is not reported. This IPS does not fulfill the requirements of infrastructure and prior data collection independence. The micro-IMUs are easily assembled on the boots and the wake-up landmarks can also be easily integrated into a commercially available smoke detectors, but the maximum operating temperature is 70^(@)C70^{\circ} \mathrm{C}, which can compromise the physical robustness of the IPS during urban fires. The handheld device is still a prototype and does not fulfill the requirements. The landmarks are battery powered, making the system fully operational even in catastrophic scenarios, and have an autonomy up to 8 years. No considerations about the power source and autonomy of the other components of the IPS are made. 响应任务。消防员的位置显示在手持设备上。至于如何向事故指挥官发送和显示这些信息，则没有报告。这种 IPS 不符合基础设施和事先数据收集独立性的要求。微型IMU 可以很容易地组装在靴子上，唤醒地标也可以很容易地集成到市面上的烟雾探测器中，但其最高工作温度为 70^(@)C70^{\circ} \mathrm{C} ，这可能会影响 IPS 在城市火灾中的物理坚固性。手持设备仍是原型，不能满足要求。地标是由电池供电的，因此即使在灾难性的情况下，系统也能完全正常运行，并且自主运行时间长达 8 年。没有考虑 IPS 其他组件的电源和自主性。
Summary of Hybrid Systems: The hybrid IPSs combine RF technologies with inertial sensors to overcome the limitations of each technological principle when used alone. Namely, they combine the continuous position estimate of the IMUbased systems with the capability of the RF signals to travel across walls for position corrections, bidirectional communication with the incident commander and immunity to the type of movement. 混合系统概述：混合式 IPS 将射频技术与惯性传感器相结合，克服了每种技术原理单独使用时的局限性。也就是说，它们将基于 IMU 的系统的连续位置估计与射频信号穿越墙壁进行位置校正的能力、与事故指挥官的双向通信以及对运动类型的抗干扰能力结合在一起。
These IPSs outperform the others in terms of accuracy and are capable of working in all environments and with all movement types. The position estimation from the RF subsystem is not affected by the movement type. Nevertheless, a better performance of the IPS will be achieved from a correct identification and adaptation to the movement type. 这些 IPS 在精度方面优于其他 IPS，并且能够在所有环境和所有运动类型下工作。射频子系统的位置估计不受运动类型的影响。不过，如果能正确识别和适应运动类型，IPS 就能获得更好的性能。
Despite all the benefits of these IPSs, the integration of several technologies increases the complexity of the data fusion algorithms. Thus, albeit the researchers intended to fuse the data from the several subsystems, some did not implement it and others implemented a simplified version of the system. Together with the increase of complexity, the development time also increases, as well as, the cost of the IPS. Furthermore, the hybrid IPSs, unlike the IMU-based ones, have a radiation signature that makes the system detectable to third parties. In some applications, like in the military, this can be critical for the security of the personnel. 尽管这些 IPSs 有很多优点，但几种技术的集成增加了数据融合算法的复杂性。因此，尽管研究人员打算融合来自多个子系统的数据，但有些人没有实施，有些人则实施了简化版系统。在复杂性增加的同时，开发时间和 IPS 的成本也随之增加。此外，混合式 IPS 与基于 IMU 的 IPS 不同，具有辐射特征，可被第三方检测到。在某些应用领域，如军事领域，这对人员安全至关重要。
D. Other Systems D.其他系统
The majority of the IPSs developed for emergency responders are based on radio technologies, inertial sensors or a combination of these two. However, due to limitations that those IPSs still have, several researchers proposed IPSs based on alternative technologies, such as artificial magnetic fields, ultrasound and security rope with communication capabilities. In this section, the IPSs for emergency responders that rely on alternative technologies are presented and discussed. 为应急响应人员开发的 IPS 大多基于无线电技术、惯性传感器或两者的结合。然而，由于这些 IPS 仍有局限性，一些研究人员提出了基于替代技术的 IPS，如人工磁场、超声波和具有通信功能的安全绳。本节将介绍和讨论依赖于替代技术的应急响应器 IPS。
The Pathfinder is a commercial system developed by Summit Safety Inc. to support Rapid Intervention Teams (RIT) in localizing firefighters that become disabled, lost, or disoriented during urban firefighting missions [116]. Pathfinder 是由 Summit Safety Inc. 开发的一种商业系统，用于支持快速干预小组 (RIT) 在执行城市消防任务时定位伤残、迷路或迷失方向的消防员[116]。
This IPS comprises the following components: a frontand a back-mounted beacon transmitter - firefighter beacon attached to the breathing gear; a tracker device carried by the 该 IPS 由以下部分组成：一个前置和一个后置的信标发射器--连接在呼吸装备上的消防员信标；一个由消防员携带的跟踪装置；一个由消防员携带的跟踪装置；一个由消防员携带的跟踪装置；一个由消防员携带的跟踪装置；一个由消防员携带的跟踪装置。
RIT; auxiliary beacons deployed as the RIT enter the building, guiding them back to the exit, and exit beacons placed at the building entrances, creating shortcuts to the path generated by the auxiliary beacons. The different beacons (auxiliary, exit, and firefighter) have different frequencies and are based on the ultrasound technology. Although the ultrasonic waves are blocked by the walls, they may find an alternative path around the corners or beneath the doors. Additionally, the ultrasonic waves are immune to smoke, heat, humidity and audible sound, making them an attractive solution in urban fires. 这些信标包括：RIT 信标；在 RIT 进入大楼时部署的辅助信标，引导他们返回出口；以及放置在大楼入口处的出口信标，为辅助信标生成的路径创建捷径。不同的信标（辅助信标、出口信标和消防员信标）具有不同的频率，并以超声波技术为基础。虽然超声波会被墙壁阻挡，但它们可以在墙角或门下找到其他路径。此外，超声波不受烟雾、热量、湿度和声音的影响，因此在城市火灾中是一种极具吸引力的解决方案。
When a firefighter is motionless or activates the panic button, the respective beacon will start transmitting an ultrasonic signal. The RIT enters the building, tracks the position of the firefighter with the tracker device and deploys the auxiliary beacons to track their way back. By switching the tracker to the exit mode, the system will guide the RIT to the nearest exit. Only the direction of the firefighter in distress is provided therefore, his/her position is not calculated. The direction of the firefighter is determined by the amplitude of the signal and is represented by a bar chart in the tracker device. This information is only accessible to the RIT. When compared with traditional systems, the search time is reduced in 50 to 80%80 \%. 当消防员一动不动或启动紧急按钮时，相应的信标将开始发射超声波信号。RIT 进入大楼后，通过跟踪装置追踪消防员的位置，并部署辅助信标以追踪其返回路线。将跟踪器切换到出口模式后，系统将引导 RIT 抵达最近的出口。系统只提供遇险消防员的方向，因此不会计算其位置。消防员的方向由信号的振幅决定，并通过跟踪装置中的条形图表示。这些信息只有 RIT 才能获取。与传统系统相比，搜索时间缩短了 50 至 80%80 \% 。
The system is independent of the building infrastructure and prior data collection. Its coverage is limited by the maximum range of the beacons ( 30 meters in LOS), which may increase the search time in complex buildings. The weight of both the firefighter’s beacon and the tracker device are compatible with the requirements specified in Section III-D. The size of the tracker device does not fulfill the requirements, but as it is carried only by the rescuer RIT it does not interfere with the firefighters activities. The firefighter’s beacons are easily assembled on the breathing gear. Both components are battery powered and have an autonomy up to 100 and 25 hours, respectively. With the firefighter’s beacons in the transmitting mode, the autonomy of the system is one hour. 该系统不受建筑物基础设施和事先数据收集的影响。其覆盖范围受到信标最大射程的限制（LOS 时为 30 米），这可能会增加在复杂建筑物中的搜索时间。消防员信标和跟踪装置的重量均符合第 III-D 节中规定的要求。跟踪装置的尺寸不符合要求，但由于它仅由救援者 RIT 携带，因此不会干扰消防员的活动。消防员信标很容易安装在呼吸装备上。这两个组件都由电池供电，使用寿命分别长达 100 小时和 25 小时。当消防员信标处于发射模式时，系统的使用寿命为一小时。
The Magnetic Indoor Local Positioning System (MILPS) is an IPS that was not developed specifically for emergency responders but was designed to operate in harsh environments, like the ones faced by the emergency responders during their missions [117]-[120]. 磁性室内局部定位系统（MILPS）是一种 IPS，它不是专门为应急响应人员开发的，而是为在恶劣环境下运行而设计的，比如应急响应人员在执行任务时所面临的环境[117]-[120]。
The system comprises a Mobile Station (MS) and three Reference Stations (RSs) deployed around the building. The MS is a smartphone equipped with a magnetic sensor and the RSs are coils powered by direct current to generate artificial magnetic fields. The distance between the MS and RS is calculated based on the intensity of the magnetic field received by the MS. The position of the MS is calculated based on the trilateration method and the distances are calculated from the several RSs whose position is known beforehand. To distinguish the different magnetic fields, these are sequentially generated using a real time clock. Differential readings and an adaptive filter are used to reduce the noise generated by other magnetic fields. 该系统包括一个移动站（MS）和三个参考站（RS），部署在大楼周围。MS 是一部配备磁传感器的智能手机，RS 是由直流电驱动的线圈，用于产生人工磁场。MS 与 RS 之间的距离根据 MS 接收到的磁场强度计算得出。MS 的位置是根据三坐标法计算出来的，而距离则是根据事先已知位置的几个 RS 计算出来的。为了区分不同的磁场，使用实时时钟依次生成这些磁场。差分读数和自适应滤波器用于减少其他磁场产生的噪音。
This IPS has an accuracy of 0.5 meters without calibration and 0.1 meters with calibration. The MS calculates and displays its position to the user. This information is not sent to other entities. The sampling rate of the IPS is 6.5 milliseconds. 该 IPS 不校准时精度为 0.5 米，校准后精度为 0.1 米。MS 会计算并向用户显示其位置。该信息不会发送给其他实体。IPS 的采样率为 6.5 毫秒。
Although independent of building infrastructure, the system relies on prior data collection, namely, the position of the RSs. Additionally, the limited range of artificial magnetic fields created by the RSs ( 15 meters) restricts the coverage of the IPS in complex buildings. No considerations are made about the modularity, physical robustness, size, weight and ease of assembling, as well as, the power source and autonomy of the system. 虽然该系统独立于建筑基础设施，但它依赖于事先的数据收集，即 RS 的位置。此外，RS 产生的人工磁场范围有限（15 米），限制了 IPS 在复杂建筑中的覆盖范围。没有考虑系统的模块化、物理坚固性、尺寸、重量和组装简便性，也没有考虑系统的电源和自主性。
The Smart LifeLine was developed under the ProFiTex project to support firefighters during urban firefighting missions [121], [122]. This IPS integrates a jacket with embedded sensors and electronics (e.g., infrared camera, IMU, and HMD) and a braided data and security rope - called “Smart LifeLine” - as a medium of data and energy transmission. 智能生命线是在 ProFiTex 项目下开发的，用于在城市消防任务中为消防员提供支持[121]，[122]。这种 IPS 集成了一件带有嵌入式传感器和电子设备（如红外摄像机、IMU 和 HMD）的外套，以及一条编织数据和安全绳（称为 "智能生命线"），作为数据和能量传输的媒介。
The Smart LifeLine is an improved version of the traditional security ropes used by the emergency responders during urban firefighting. By adding data transmission capabilities it provides an innovative method for data transmission between the firefighters and the incident commander. Several beacons are embedded within the rope, enabling navigation and localization of firefighters in smoky environments and data exchange between the firefighters and the incident commander. Additionally, the system allows monitoring the firefighter’s condition (e.g., movement pattern, posture, heart rate, and environmental temperature) and triggering alarms if problems are detected. The system is mentioned to have localization capability, but the localization accuracy is not reported. An HMD is used to display the distance to the nearest beacon and to the exit, as well as, the other monitored parameters. The data obtained from the infrared camera is combined with a SLAM algorithm for a 3D building reconstruction and thermal mapping as the firefighter enters a building. In the current implementation, the SLAM algorithm runs offline. 智能生命线是城市消防应急人员使用的传统安全绳索的改进版。通过增加数据传输功能，它为消防员和事故指挥官之间的数据传输提供了一种创新方法。绳索中嵌入了多个信标，可在烟雾弥漫的环境中导航和定位消防员，并在消防员和事故指挥官之间进行数据交换。此外，该系统还能监测消防员的状况（如运动模式、姿势、心率和环境温度），并在发现问题时触发警报。系统具有定位功能，但没有报告定位精度。HMD 用于显示与最近的信标和出口的距离，以及其他监测参数。从红外摄像机获得的数据与 SLAM 算法相结合，可在消防员进入建筑物时进行三维建筑物重建和热映射。在目前的实施中，SLAM 算法是离线运行的。
This IPS is independent of the building infrastructure and prior data collection. The coverage is limited by the length of the rope and may be reduced if it gets stuck in the furniture, doors or other building structures. No considerations are made about the modularity, physical robustness, size, weight and ease of assembling, as well as, the power source and autonomy of the system. 这种 IPS 不受建筑基础设施和事先数据收集的影响。覆盖范围受到绳索长度的限制，如果绳索被家具、门或其他建筑结构卡住，覆盖范围可能会缩小。系统的模块化、物理坚固性、尺寸、重量和组装简便性以及电源和自主性均不在考虑之列。
Summary of Other Systems: In this subsection, some alternative approaches addressing the indoor localization problem of emergency responders were introduced and discussed. The main limitation of the ultrasound-based IPSs is due to the inability of ultrasonic waves to travel through walls. Therefore, in a complex building, the number of beacon nodes has to be high to cover the entire building, which is unfeasible during emergency responders’ missions. A promising approach is based on the generation of artificial magnetic fields. The magnetic fields are immune to the multipath and fading phenomena and can easily penetrate obstacles, achieving a high performance, even in NLOS conditions. Nevertheless, they may be distorted by the building materials and their range is short. The use of security ropes with communication capabilities present the same problems of traditional security ropes, i.e., they may get stuck on furniture, door, or other building structures, reducing the coverage of the system. 其他系统概述：本小节介绍并讨论了解决应急响应人员室内定位问题的一些替代方法。基于超声波的 IPS 的主要局限性在于超声波无法穿墙。因此，在复杂的建筑物中，信标节点的数量必须很多才能覆盖整个建筑物，这在应急响应人员执行任务时是不可行的。一种很有前景的方法是基于人工磁场的生成。人工磁场不受多径和衰减现象的影响，可以轻易穿透障碍物，即使在 NLOS 条件下也能实现高性能。不过，这些磁场可能会被建筑材料扭曲，而且传播距离较短。使用具有通信功能的安全绳索与传统安全绳索存在同样的问题，即它们可能会卡在家具、门或其他建筑结构上，从而降低系统的覆盖范围。
E. Comparison of IPS Schemes for Emergency Responders E.应急响应人员 IPS 方案比较
In Table II, Table III and Table IV, the IPSs surveyed in the previous subsections are evaluated and compared from the viewpoints introduced in Section V. In the three tables, the IPSs are grouped by the technological principle adopted and are referenced by system/author name, year and a reference to the respective paper. 在表 II、表 III 和表 IV 中，根据第 V 部分介绍的观点对前面各小节调查的 IPS 进行了评估和比较。在这三个表格中，IPS 按照所采用的技术原理分组，并以系统/作者姓名、年份和相关论文的参考文献作为参考。
In Table II, the IPSs are compared with respect to the main design choices, namely, technological principle, deployment, localization principle, algorithm and environment. Additionally, the column Technologies lists the technology types used by each system and, the column Techniques shows the different techniques used in each IPS for the localization process. 表 II 比较了 IPS 的主要设计选择，即技术原理、部署、定位原理、算法和环境。此外，"技术 "一栏列出了每个系统使用的技术类型，"技术 "一栏显示了每个 IPS 在定位过程中使用的不同技术。
Table III compares the IPSs regarding the requirements of an IPS for emergency responders, namely, accuracy, information accessibility, adaptability, architecture, autonomy and cost. Concerning the requirements of information accessibility, adaptability and architecture, the symbols " ✓\checkmark " and “X” are used to represent that the IPS complies or does not comply with the requirement, respectively. The symbol “-” is used when no information about a requirement is provided. The column Cost represents the deployment expenses of each system and is classified as low (L), medium (M) or high (H). This categorization is based on the price of the used sensors and the number of components needed since, in most cases, the cost of the IPS is not provided. 表 III 比较了 IPS 对应急响应者的要求，即准确性、信息可获取性、适应性、结构、自 主性和成本。关于信息无障碍、适应性和结构的要求，符号" ✓\checkmark "和 "X "分别表示 IPS 符合或不符合要求。符号"-"表示没有提供有关要求的信息。成本一栏表示每个系统的部署费用，分为低（L）、中（M）或高（H）。这种分类是基于所用传感器的价格和所需组件的数量，因为在大多数情况下，没有提供 IPS 的成本。
In Table IV, the IPSs are compared in relation to some additional features of each IPS. The column Layout indicates the dependence of the IPS from the building plan. The major limitations of each IPS are described in the column Limitations. The column Other Features lists the functionalities that each IPS has besides localization. 表 IV 比较了 IPS 的一些附加功能。布局 "一栏说明了 IPS 对建筑平面图的依赖程度。限制 "一栏说明了每个 IPS 的主要限制。其他功能一栏列出了每个 IPS 除定位之外的其他功能。
The information provided in these comparative tables can be used as a tool for the design of future IPSs for emergency responders, as they enable finding the best design choices that fulfill the IPS requirements. 这些比较表中提供的信息可作为设计未来应急响应人员 IPS 的工具，因为它们有助于找到满足 IPS 要求的最佳设计方案。
VII. Issues, Challenges, and Future Research Directions VII.问题、挑战和未来研究方向
Several IPSs have been developed to address the problem of indoor localization in scenarios faced by emergency responders. In fact, harsh conditions and unstructured environments raise a lot of challenges in the design of an IPS for emergency responders. In the following subsections, issues, challenges and future research directions in the design of an IPS for emergency responders will be highlighted. 为了解决应急响应人员所面临的室内定位问题，已经开发了几种 IPS。事实上，恶劣的条件和非结构化环境给应急响应人员 IPS 的设计带来了很多挑战。以下各小节将重点介绍为应急响应人员设计 IPS 的问题、挑战和未来研究方向。
A. Issues and Challenges in IPSs for Emergency Responders A.应急响应人员 IPS 的问题和挑战
Many studies have been carried out on indoor localization, but none of the developed systems is capable of meeting all the requirements of emergency responders for an IPS yet. In part, as depicted by Harris, the main reason for the failure is the lack of cooperation between the IPS development teams and the emergency responders [25]. This lack of cooperation implies that the majority of the developed systems are unable to adapt 关于室内定位的研究已经开展了很多，但所开发的系统还没有一个能够满足应急响应者对 IPS 的所有要求。Harris 指出，失败的部分主要原因是 IPS 开发团队与应急响应人员之间缺乏合作[25]。这种合作的缺失意味着大多数已开发的系统都无法适应应急响应者对 IPS 的所有要求。
TABLE II 表 II
Comparison of the Surveyed IPSs Based on Design Choices 基于设计选择的调查 IPS 比较
System/Author Name Year Technologies Techniques Deployment Localization Principle Algorithm Environment
Radio Signal-Based Systems
SmokeNet-2007 [65] ZigBee; RF Signal RSSI; Fingerprinting Preinstalled WSN Decentralized Indoor
" Europcom - 2008 [17],
[66], [67] " UWB; Wi-Fi RToF; Iterative Least Squares Strategic Ad-Hoc Decentralized Indoor
ProeTEX - 2009 [68] GPS Trilateration No Infrastructure-Based Decentralized Outdoor
LifeNet - 2009 [70] Ultrasound; RF TDoA; RSSI Strategic Ad-Hoc Decentralized Indoor
FIREGUIDE - 2010 [71] RFID; Bluetooth; Wi-Fi Response Rate; Proximity Preinstalled Infrastructure-Based and Proximity Decentralized Indoor
Ruppel el al. - 2010 [72] RFID; Bluetooth; Wi-Fi Multi-Method-Approach (MMA); Fingerprinting; BIM; Trilateration; Proximity Preinstalled Infrastructure-Based and Proximity Centralized Indoor
"Li and Gerber-2012" Wi-Fi RSSI; Triangulation; Propagation Model No Ad-Hoc Decentralized Indoor
Moon et al. - 2013 [74], [75] RF; GPS SLAM; EKF; ToF Strategic Ad-Hoc Centralized Indoor and Outdoor
Ghosh et al. - 2013 [76] ZigBee RSSI Strategic Ad-Hoc Centralized Indoor
Zhang et al. - 2013 [53] UWB TDoA; AoA; EKF; MultiDimensional Markov Jump System Strategic and Preinstalled Ad-Hoc Centralized Indoor
Li et al. - 2014 [77] Wi-Fi BIM; SBL; RSSI; Centroid; MLE Strategic and Preinstalled Infrastructure-Based and Ad-Hoc Centralized Indoor
Femminella and Reali - 2015" [18] " Wi-Fi; GPS Trilateration; RSSI Strategic Ad-Hoc Decentralized Indoor and Outdoor
IMU-Based Systems
Beauregard - 2006 [80] IMU; GPS PDR; Feedforward Neural Network; Trilateration No Dead Reckoning Decentralized Indoor and Outdoor
Beauregard - 2007 [81] IMU Dead Reckoning; Strap-Down Equations; ZUTP No Dead Reckoning Decentralized Indoor and Outdoor
IndoorNav - 2007 [82] IMU; RFID PDR; Map Matching; EKF; Proximity Strategic Dead Reckoning and Proximity Centralized Indoor
Ojeda and Borenstein - 2007" [83] " IMU PDR; ZUPT No Dead Reckoning Decentralized Indoor
Widyawan et al. - 2008 [84] IMU PDR; Particle Filter; Map Matching No Dead Reckoning Centralized Indoor
HeadSLAM - 2008 [86] IMU; Laser Scanner PDR; Particle Filter; SLAM No Dead Reckoning Decentralized Indoor
CADMS - 2010 [89] IMU PDR; EKF; ZUPT; Map Matching No Dead Reckoning Centralized Indoor
Rantakokko et al. - 2011 [90] IMU; UWB; GPS PDR; ZUPT; KF; Cooperative Localization; Trilateration No Dead Reckoning Centralized Indoor and Outdoor
Zhang et al. - 2012 [92] IMU PDR; ZUPT; KF No Dead Reckoning Decentralized Indoor
Hari el al. - 2013 [94] 2 IMUs; UWB; Camera PDR; ZUPT; KF; Centralized Cooperative Localization Preinstalled Dead Reckoning and Infrastructure-Based Centralized Indoor
REFIRE - 2013 [13] IMU; RFID PDR; EKF; Online learning algorithm Preinstalled Dead Reckoning and Infrastructure-Based Decentralized Indoor
Rantakokko et al. - 2014 [19] 2 IMUs PDR; ZUPT; constrained EKF No Dead Reckoning Decentralized Indoor
TOR - 2014 [100] 2 IMUs; UWB PDR; ZUPT; KF; Centralized Cooperative Localization; Synthetic Ranging Preinstalled Dead Reckoning and Infrastructure-Based Centralized Indoor
Hybrid Systems
" Relate Trails - 2008 [21],
[101], [102] " IMU; Ultrasound PDR; ZUPT; Proximity Strategic Dead Reckoning and Proximity Decentralized Indoor
PPL-2011 [103]-[107] Radio Signals; IMU KF; Trilateration Strategic Dead Reckoning and Ad Hoc Centralized Indoor
" Virtual Lifeline - 2011
[108] " IMU; Ultrasound; RF (ZigBee) PDR; ZUPT; Proximity; RSSI: KF; PF; Backtracking PF Strategic and Preinstalled Dead Reckoning, Ad Hoc, and Proximity Decentralized Indoor
GLANSER - 2012 [109] 6DoF IMU; Doppler; Barometer, Motion Model; UWB; GPS EKF; Synthetic Aperture; Trilateration Strategic Dead Reckoning and Ad Hoc Decentralized Indoor and Outdoor
WASP - 2013 [112] IMU; Pressure Sensor; Ranging Sensors; NFC; GPS ToF; Map Matching; Triangulation; Patented sensor fusion algorithms Strategic and Preinstalled Dead Reckoning, Infrastructure-Based, and Ad Hoc Decentralized Indoor and Outdoor
Simon el al. 2015 [114] RF; IMU RSSI; Gradient Descent method; Gauss-Newton Algorithm; PDR; ZUPT; KF Preinstalled Dead Reckoning and Infrastructure-Based Decentralized Indoor
Other Systems
PathFinder [116] Ultrasound RSSI Strategic Ad Hoc Decentralized Indoor
MILPS [117]-[120] Artificial Magnetic Fields Trilateration; Differential Measurement; Adaptive Filter Strategic Ad Hoc Decentralized Indoor
" Smart Lifeline - 2013
[121], [122] " IMU; Infrared Camera; Depth Sensor SLAM Strategic Infrastructure-Based Centralized Indoor| System/Author Name Year | Technologies | Techniques | Deployment | Localization Principle | Algorithm | Environment |
| :---: | :---: | :---: | :---: | :---: | :---: | :---: |
| Radio Signal-Based Systems | | | | | | |
| SmokeNet-2007 [65] | ZigBee; RF Signal | RSSI; Fingerprinting | Preinstalled | WSN | Decentralized | Indoor |
| $\begin{aligned} & \text { Europcom - } 2008 \text { [17], } \\ & \text { [66], [67] } \end{aligned}$ | UWB; Wi-Fi | RToF; Iterative Least Squares | Strategic | Ad-Hoc | Decentralized | Indoor |
| ProeTEX - 2009 [68] | GPS | Trilateration | No | Infrastructure-Based | Decentralized | Outdoor |
| LifeNet - 2009 [70] | Ultrasound; RF | TDoA; RSSI | Strategic | Ad-Hoc | Decentralized | Indoor |
| FIREGUIDE - 2010 [71] | RFID; Bluetooth; Wi-Fi | Response Rate; Proximity | Preinstalled | Infrastructure-Based and Proximity | Decentralized | Indoor |
| Ruppel el al. - 2010 [72] | RFID; Bluetooth; Wi-Fi | Multi-Method-Approach (MMA); Fingerprinting; BIM; Trilateration; Proximity | Preinstalled | Infrastructure-Based and Proximity | Centralized | Indoor |
| Li and Gerber-2012 | Wi-Fi | RSSI; Triangulation; Propagation Model | No | Ad-Hoc | Decentralized | Indoor |
| Moon et al. - 2013 [74], [75] | RF; GPS | SLAM; EKF; ToF | Strategic | Ad-Hoc | Centralized | Indoor and Outdoor |
| Ghosh et al. - 2013 [76] | ZigBee | RSSI | Strategic | Ad-Hoc | Centralized | Indoor |
| Zhang et al. - 2013 [53] | UWB | TDoA; AoA; EKF; MultiDimensional Markov Jump System | Strategic and Preinstalled | Ad-Hoc | Centralized | Indoor |
| Li et al. - 2014 [77] | Wi-Fi | BIM; SBL; RSSI; Centroid; MLE | Strategic and Preinstalled | Infrastructure-Based and Ad-Hoc | Centralized | Indoor |
| Femminella and Reali - $2015 \text { [18] }$ | Wi-Fi; GPS | Trilateration; RSSI | Strategic | Ad-Hoc | Decentralized | Indoor and Outdoor |
| IMU-Based Systems | | | | | | |
| Beauregard - 2006 [80] | IMU; GPS | PDR; Feedforward Neural Network; Trilateration | No | Dead Reckoning | Decentralized | Indoor and Outdoor |
| Beauregard - 2007 [81] | IMU | Dead Reckoning; Strap-Down Equations; ZUTP | No | Dead Reckoning | Decentralized | Indoor and Outdoor |
| IndoorNav - 2007 [82] | IMU; RFID | PDR; Map Matching; EKF; Proximity | Strategic | Dead Reckoning and Proximity | Centralized | Indoor |
| Ojeda and Borenstein - $2007 \text { [83] }$ | IMU | PDR; ZUPT | No | Dead Reckoning | Decentralized | Indoor |
| Widyawan et al. - 2008 [84] | IMU | PDR; Particle Filter; Map Matching | No | Dead Reckoning | Centralized | Indoor |
| HeadSLAM - 2008 [86] | IMU; Laser Scanner | PDR; Particle Filter; SLAM | No | Dead Reckoning | Decentralized | Indoor |
| CADMS - 2010 [89] | IMU | PDR; EKF; ZUPT; Map Matching | No | Dead Reckoning | Centralized | Indoor |
| Rantakokko et al. - 2011 [90] | IMU; UWB; GPS | PDR; ZUPT; KF; Cooperative Localization; Trilateration | No | Dead Reckoning | Centralized | Indoor and Outdoor |
| Zhang et al. - 2012 [92] | IMU | PDR; ZUPT; KF | No | Dead Reckoning | Decentralized | Indoor |
| Hari el al. - 2013 [94] | 2 IMUs; UWB; Camera | PDR; ZUPT; KF; Centralized Cooperative Localization | Preinstalled | Dead Reckoning and Infrastructure-Based | Centralized | Indoor |
| REFIRE - 2013 [13] | IMU; RFID | PDR; EKF; Online learning algorithm | Preinstalled | Dead Reckoning and Infrastructure-Based | Decentralized | Indoor |
| Rantakokko et al. - 2014 [19] | 2 IMUs | PDR; ZUPT; constrained EKF | No | Dead Reckoning | Decentralized | Indoor |
| TOR - 2014 [100] | 2 IMUs; UWB | PDR; ZUPT; KF; Centralized Cooperative Localization; Synthetic Ranging | Preinstalled | Dead Reckoning and Infrastructure-Based | Centralized | Indoor |
| Hybrid Systems | | | | | | |
| $\begin{aligned} & \text { Relate Trails - } 2008 \text { [21], } \\ & \text { [101], [102] } \end{aligned}$ | IMU; Ultrasound | PDR; ZUPT; Proximity | Strategic | Dead Reckoning and Proximity | Decentralized | Indoor |
| PPL-2011 [103]-[107] | Radio Signals; IMU | KF; Trilateration | Strategic | Dead Reckoning and Ad Hoc | Centralized | Indoor |
| $\begin{aligned} & \text { Virtual Lifeline - } 2011 \\ & \text { [108] } \end{aligned}$ | IMU; Ultrasound; RF (ZigBee) | PDR; ZUPT; Proximity; RSSI: KF; PF; Backtracking PF | Strategic and Preinstalled | Dead Reckoning, Ad Hoc, and Proximity | Decentralized | Indoor |
| GLANSER - 2012 [109] | 6DoF IMU; Doppler; Barometer, Motion Model; UWB; GPS | EKF; Synthetic Aperture; Trilateration | Strategic | Dead Reckoning and Ad Hoc | Decentralized | Indoor and Outdoor |
| WASP - 2013 [112] | IMU; Pressure Sensor; Ranging Sensors; NFC; GPS | ToF; Map Matching; Triangulation; Patented sensor fusion algorithms | Strategic and Preinstalled | Dead Reckoning, Infrastructure-Based, and Ad Hoc | Decentralized | Indoor and Outdoor |
| Simon el al. 2015 [114] | RF; IMU | RSSI; Gradient Descent method; Gauss-Newton Algorithm; PDR; ZUPT; KF | Preinstalled | Dead Reckoning and Infrastructure-Based | Decentralized | Indoor |
| Other Systems | | | | | | |
| PathFinder [116] | Ultrasound | RSSI | Strategic | Ad Hoc | Decentralized | Indoor |
| MILPS [117]-[120] | Artificial Magnetic Fields | Trilateration; Differential Measurement; Adaptive Filter | Strategic | Ad Hoc | Decentralized | Indoor |
| $\begin{aligned} & \text { Smart Lifeline - } 2013 \\ & \text { [121], [122] } \end{aligned}$ | IMU; Infrared Camera; Depth Sensor | SLAM | Strategic | Infrastructure-Based | Centralized | Indoor |
to the dynamic scenarios faced by emergency responders during their on-duty missions, resulting in a poor performance of the system in terms of accuracy and precision. In this 这就导致系统在准确性和精确性方面表现不佳。在这种情况下
subsection, the major issues and challenges in the design and performance assessment of IPSs for emergency responders are discussed. 小节讨论了为应急响应人员设计和评估 IPS 性能的主要问题和挑战。
TABLE III 表 III
Comparison of the Surveyed IPSs Based on the IPS Requirements of Emergency Responders 根据应急响应人员对 IPS 的要求对所调查的 IPS 进行比较
System/Author Name -Year Accuracy Information Accessibility Adaptability of the IPS Architecture of the IPS Autonomy Cost
A1 A2 A3 A4 A5 B1 B2 B3 B4 B5 C1 C2 C3 C4 C5
Radio Signal-Based Systems
SmokeNet-2007 [65] Room Level ✓ ✓ ✓ X - X X X X ✓ - X - - - 95 hours H
" Europcom - 2008 [17],
[66], [67] " 1 m X ✓ - X - ✓ ✓ ✓ ✓ - - ✓ - - X - H
ProeTEX - 2009 [68] - X ✓ ✓ X - ✓ ✓ ✓ ✓ - ✓ X ✓ ✓ ✓ 7 hours L
LifeNet - 2009 [70] "Relative
Position" ✓ ✓ - X - ✓ ✓ ✓ ✓ ✓ X ✓ X X X - M
" FIREGUIDE - 2010
[71] " "Room
Level" ✓ X - X - X X X X ✓ X X - ✓ ✓ Powered by the grid L
" Ruppel el al. -2010
[72] " - ✓ ✓ - X - X X X X ✓ X ✓ - X X Powered by the grid H
" Li and Gerber- 2012
[20] " 2,24m (only in a room) X X ✓ X - ✓ X ✓ X ✓ X - - ✓ ✓ - L
Moon et al. - 2013 [74], [75] <1m X X - ✓ - ✓ ✓ ✓ ✓ - - X - - - - M
Ghosh et al. - 2013 [76] - X X ✓ X - ✓ ✓ X ✓ - - ✓ - - - 2 months L
Zhang et al. - 2013 [53] "0,03m (in
LOS) " X ✓ X X ✓ ✓ ✓ X ✓ X - X X - - - H
Li et al. - 2014 [77] Room level (>82.8%); coordinatelevel (< 2.29 m in 95%) X ✓ ✓ X - ✓ X ✓ ✓ ✓ X ✓ ✓ - ✓ - L
Femminella and Reali- 2015" [18] " < 10m X ✓ ✓ X - ✓ ✓ X X - X ✓ - - - - M
IMU-Based Systems
Beauregard - 2006 [80] - X X X X ✓ ✓ ✓ ✓ ✓ ✓ X - X ✓ X - L
Beauregard - 2007 [81] - X X X X ✓ ✓ ✓ ✓ ✓ ✓ ✓ - X X X - L
IndoorNav - 2007 [82] " < 5m in
90%" X X - X ✓ ✓ X X X ✓ ✓ - X X X - M
Ojeda and Borenstein 2007 [83] < 2,1% of distance traveled X X X X ✓ ✓ ✓ ✓ ✓ ✓ ✓ - X X X - L
Widyawan et al. - 2008 [84] "1.89-
1.55m" X X X X ✓ ✓ ✓ X ✓ ✓ X - X X X - L
HeadSLAM - 2008 [86] 12 m X X X ✓ - ✓ ✓ ✓ ✓ ✓ X - - X X - L
CADMS - 2010 [89] " < 1,7m in
95%" ✓ ✓ - X - ✓ X X - ✓ X - X X X - L
Rantakokko et al. 2011 [90] < 3.5m X X - X - ✓ ✓ - ✓ ✓ ✓ - X X X - M
Zhang et al. - 2012 [92] "< 1,11 m for walking; < 7,73m (in Z
direction) for running" X X - X ✓ ✓ ✓ ✓ ✓ ✓ X ✓ X X X - L
Hari el al. - 2013 [94] - X X - X X X ✓ X X ✓ ✓ ✓ ✓ ✓ ✓ 1.5 hours M
REFIRE-2013 [13] "Room
Level" ✓ - X X - X X X X ✓ ✓ ✓ X - - - M
Rantakokko et al. 2014 [19] <2m X X X X ✓ ✓ ✓ ✓ ✓ ✓ X - X X X - L
TOR-2014 [100] < 3m X X ✓ X - X ✓ X X ✓ X ✓ X X ✓ 1.5 hours M
Hybrid Systems
" Relate Trails - 2008
[21], [101], [102] " Room Level ✓ ✓ - ✓ - ✓ ✓ ✓ ✓ ✓ X ✓ X X X - M
"PPL-2011[103]
[107]" 0.14 m X ✓ ✓ X ✓ ✓ ✓ ✓ X X - ✓ - ✓ ✓ 24 hours M
" Virtual Lifeline - 2011
[108] " <8,66m X X - X - ✓ ✓ ✓ ✓ - X ✓ X X X - M
GLANSER - 2012 [109] 3 m (only sensors) X ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ - X ✓ - ✓ X - H
WASP - 2013 [112] "Room
Level" ✓ ✓ ✓ ✓ ✓ ✓ X ✓ ✓ - ✓ ✓ - ✓ - - H
Simon el al. 2015 [114] "0,99m
(with
landmarks) " X - X X - X X X ✓ X ✓ - X ✓ ✓ 8 years M
Other Systems
PathFinder [116] - X ✓ ✓ X ✓ ✓ ✓ ✓ ✓ - X ✓ - ✓ X 25 hours L
MILPS [117]-[120] <0,5m X - ✓ X ✓ ✓ ✓ ✓ X - X - - - - - L
" Smart Lifeline - 2013
[121], [122] " Relative Position ✓ ✓ X ✓ ✓ ✓ ✓ X X X - X - - - - H| System/Author Name -Year | Accuracy | Information Accessibility | | | | | Adaptability of the IPS | | | | | Architecture of the IPS | | | | | Autonomy | Cost |
| :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: |
| | | A1 | A2 | A3 | A4 | A5 | B1 | B2 | B3 | B4 | B5 | C1 | C2 | C3 | C4 | C5 | | |
| Radio Signal-Based Systems | | | | | | | | | | | | | | | | | | |
| SmokeNet-2007 [65] | Room Level | $\checkmark$ | $\checkmark$ | $\checkmark$ | X | - | X | X | X | X | $\checkmark$ | - | X | - | - | - | 95 hours | H |
| $\begin{aligned} & \text { Europcom - } 2008 \text { [17], } \\ & \text { [66], [67] } \end{aligned}$ | 1 m | X | $\checkmark$ | - | X | - | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | - | - | $\checkmark$ | - | - | X | - | H |
| ProeTEX - 2009 [68] | - | X | $\checkmark$ | $\checkmark$ | X | - | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | - | $\checkmark$ | X | $\checkmark$ | $\checkmark$ | $\checkmark$ | 7 hours | L |
| LifeNet - 2009 [70] | Relative <br> Position | $\checkmark$ | $\checkmark$ | - | X | - | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | X | $\checkmark$ | X | X | X | - | M |
| $\begin{aligned} & \text { FIREGUIDE - } 2010 \\ & \text { [71] } \end{aligned}$ | Room <br> Level | $\checkmark$ | X | - | X | - | X | X | X | X | $\checkmark$ | X | X | - | $\checkmark$ | $\checkmark$ | Powered by the grid | L |
| $\begin{aligned} & \text { Ruppel el al. }-2010 \\ & \text { [72] } \end{aligned}$ | - | $\checkmark$ | $\checkmark$ | - | X | - | X | X | X | X | $\checkmark$ | X | $\checkmark$ | - | X | X | Powered by the grid | H |
| $\begin{aligned} & \text { Li and Gerber- } 2012 \\ & \text { [20] } \end{aligned}$ | 2,24m (only in a room) | X | X | $\checkmark$ | X | - | $\checkmark$ | X | $\checkmark$ | X | $\checkmark$ | X | - | - | $\checkmark$ | $\checkmark$ | - | L |
| Moon et al. - 2013 [74], [75] | <1m | X | X | - | $\checkmark$ | - | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | - | - | X | - | - | - | - | M |
| Ghosh et al. - 2013 [76] | - | X | X | $\checkmark$ | X | - | $\checkmark$ | $\checkmark$ | X | $\checkmark$ | - | - | $\checkmark$ | - | - | - | 2 months | L |
| Zhang et al. - 2013 [53] | $\begin{aligned} & 0,03 \mathrm{~m} \text { (in } \\ & \text { LOS) } \end{aligned}$ | X | $\checkmark$ | X | X | $\checkmark$ | $\checkmark$ | $\checkmark$ | X | $\checkmark$ | X | - | X | X | - | - | - | H |
| Li et al. - 2014 [77] | Room level (>82.8%); coordinatelevel (< 2.29 m in 95%) | X | $\checkmark$ | $\checkmark$ | X | - | $\checkmark$ | X | $\checkmark$ | $\checkmark$ | $\checkmark$ | X | $\checkmark$ | $\checkmark$ | - | $\checkmark$ | - | L |
| Femminella and Reali- $2015 \text { [18] }$ | $<10 \mathrm{~m}$ | X | $\checkmark$ | $\checkmark$ | X | - | $\checkmark$ | $\checkmark$ | X | X | - | X | $\checkmark$ | - | - | - | - | M |
| IMU-Based Systems | | | | | | | | | | | | | | | | | | |
| Beauregard - 2006 [80] | - | X | X | X | X | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | X | - | X | $\checkmark$ | X | - | L |
| Beauregard - 2007 [81] | - | X | X | X | X | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | - | X | X | X | - | L |
| IndoorNav - 2007 [82] | $\begin{aligned} & <5 \mathrm{~m} \text { in } \\ & 90 \% \end{aligned}$ | X | X | - | X | $\checkmark$ | $\checkmark$ | X | X | X | $\checkmark$ | $\checkmark$ | - | X | X | X | - | M |
| Ojeda and Borenstein 2007 [83] | $<2,1 \%$ of distance traveled | X | X | X | X | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | - | X | X | X | - | L |
| Widyawan et al. - 2008 [84] | $\begin{aligned} & 1.89- \\ & 1.55 \mathrm{~m} \end{aligned}$ | X | X | X | X | $\checkmark$ | $\checkmark$ | $\checkmark$ | X | $\checkmark$ | $\checkmark$ | X | - | X | X | X | - | L |
| HeadSLAM - 2008 [86] | 12 m | X | X | X | $\checkmark$ | - | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | X | - | - | X | X | - | L |
| CADMS - 2010 [89] | $\begin{aligned} & <1,7 \mathrm{~m} \text { in } \\ & 95 \% \end{aligned}$ | $\checkmark$ | $\checkmark$ | - | X | - | $\checkmark$ | X | X | - | $\checkmark$ | X | - | X | X | X | - | L |
| Rantakokko et al. 2011 [90] | $<3.5 \mathrm{~m}$ | X | X | - | X | - | $\checkmark$ | $\checkmark$ | - | $\checkmark$ | $\checkmark$ | $\checkmark$ | - | X | X | X | - | M |
| Zhang et al. - 2012 [92] | $<1,11$ m for walking; $<7,73 \mathrm{~m}$ (in Z <br> direction) for running | X | X | - | X | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | X | $\checkmark$ | X | X | X | - | L |
| Hari el al. - 2013 [94] | - | X | X | - | X | X | X | $\checkmark$ | X | X | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | 1.5 hours | M |
| REFIRE-2013 [13] | Room <br> Level | $\checkmark$ | - | X | X | - | X | X | X | X | $\checkmark$ | $\checkmark$ | $\checkmark$ | X | - | - | - | M |
| Rantakokko et al. 2014 [19] | <2m | X | X | X | X | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | X | - | X | X | X | - | L |
| TOR-2014 [100] | $<3 \mathrm{~m}$ | X | X | $\checkmark$ | X | - | X | $\checkmark$ | X | X | $\checkmark$ | X | $\checkmark$ | X | X | $\checkmark$ | 1.5 hours | M |
| Hybrid Systems | | | | | | | | | | | | | | | | | | |
| $\begin{aligned} & \text { Relate Trails - } 2008 \\ & \text { [21], [101], [102] } \end{aligned}$ | Room Level | $\checkmark$ | $\checkmark$ | - | $\checkmark$ | - | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | X | $\checkmark$ | X | X | X | - | M |
| $\begin{aligned} & P P L-2011[103] \\ & {[107]} \end{aligned}$ | 0.14 m | X | $\checkmark$ | $\checkmark$ | X | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | X | X | - | $\checkmark$ | - | $\checkmark$ | $\checkmark$ | 24 hours | M |
| $\begin{aligned} & \text { Virtual Lifeline - } 2011 \\ & \text { [108] } \end{aligned}$ | <8,66m | X | X | - | X | - | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | - | X | $\checkmark$ | X | X | X | - | M |
| GLANSER - 2012 [109] | 3 m (only sensors) | X | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | - | X | $\checkmark$ | - | $\checkmark$ | X | - | H |
| WASP - 2013 [112] | Room <br> Level | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | X | $\checkmark$ | $\checkmark$ | - | $\checkmark$ | $\checkmark$ | - | $\checkmark$ | - | - | H |
| Simon el al. 2015 [114] | $\begin{aligned} & 0,99 \mathrm{~m} \\ & \text { (with } \\ & \text { landmarks) } \end{aligned}$ | X | - | X | X | - | X | X | X | $\checkmark$ | X | $\checkmark$ | - | X | $\checkmark$ | $\checkmark$ | 8 years | M |
| Other Systems | | | | | | | | | | | | | | | | | | |
| PathFinder [116] | - | X | $\checkmark$ | $\checkmark$ | X | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | - | X | $\checkmark$ | - | $\checkmark$ | X | 25 hours | L |
| MILPS [117]-[120] | <0,5m | X | - | $\checkmark$ | X | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | X | - | X | - | - | - | - | - | L |
| $\begin{aligned} & \text { Smart Lifeline - } 2013 \\ & \text { [121], [122] } \end{aligned}$ | Relative Position | $\checkmark$ | $\checkmark$ | X | $\checkmark$ | $\checkmark$ | $\checkmark$ | $\checkmark$ | X | X | X | - | X | - | - | - | - | H |
A1 - Accessibility of the Position Information; A2 - Intuitive Presentation of the Positioning Data; A3 - Position Update Rate; A4 - SLAM; A5 - Information Security. A1 - 定位信息的可访问性；A2 - 定位数据的直观显示；A3 - 定位更新率；A4 - SLAM；A5 - 信息安全。
B1 - Independence of Building Infrastructure; B2 - Independence of Prior Data Collection; B3 - Adaptability to Dynamic Environment Changes; B4 Coverage; B5 - Deployment Effort; B1 - 建筑基础设施的独立性；B2 - 事先数据收集的独立性；B3 - 对动态环境变化的适应性；B4 覆盖范围；B5 - 部署工作；
High System Cost: In outdoor localization scenarios, the combination of GPS with cheap IMUs is enough to provide a navigation system that meets the user 系统成本高：在室外定位场景中，GPS 与廉价 IMU 的组合足以提供满足用户需求的导航系统。
requirements - availability, cost, scalability, accuracy and precision. However, for the indoor localization scenarios, the use of mature technologies has shown to be insufficient to solve the 这些要求包括可用性、成本、可扩展性、准确性和精确度。然而，在室内定位场景中，使用成熟的技术不足以解决以下问题
TABLE IV 表 IV
Comparison of the Surveyed IPSs Based on Additional Features 基于附加功能的受调查 IPS 比较
System/Author Name Year 系统/作者姓名 年份
Layout 布局
Limitations 局限性
Other Features 其他功能
Radio Signal-Based Systems 基于无线电信号的系统
SmokeNet-2007 [65]
Yes 是
Sensitive to changes in the environment; Requires calibration 对环境变化敏感；需要校准
Monitoring physiological and environmental parameters 监测生理和环境参数
LifeNet-2009 [70] 生命网-2009 [70]
No 没有
The beacons can be moved or destroyed; The position's information is relative. Short range communication (2,5m)(2,5 \mathrm{~m}) 信标可以移动或销毁；位置信息是相对的。短程通信 (2,5m)(2,5 \mathrm{~m})
Outside communication. Monitoring physiological and environmental parameters. Assisted navigation 外部通信。监测生理和环境参数辅助导航
FIREGUIDE - 2010 [71] 消防指南 - 2010 [71]
Yes 是
Sensitive to changes in the environment; Depends on building power 对环境变化敏感；取决于建筑动力
Indication of the nearest exit; Outside communication 指示最近的出口；外部通信
Ruppel el al. - 2010 [72] Ruppel el al.- 2010 [72]
Yes 是
Limited tests; The low availability of BIM models; Expensive site survey; Relies on building power 测试有限；BIM 模型可用性低；现场勘测费用高昂；依赖建筑动力
Records the firefighters' operation 记录消防员的行动
Li and Gerber - 2012 李和格伯 - 2012[20][20]
No 没有
Room-scale tested with LOS conditions; Requires absolute position; The propagation model parameters are site specific; The App needs to be installed on the smartphone's occupants 在 LOS 条件下进行房间尺度测试；需要绝对位置；传播模型参数取决于具体地点；应用程序需要安装在智能手机使用者身上
Localization of both first responders and occupants 确定急救人员和乘客的位置
Monitoring of physiological and environmental parameters; Prediction of the fire expansion 监测生理和环境参数；预测火势蔓延情况
Zhang et al. - 2013 [53] Zhang et al.
Optional 可选
Requires cables and a calibration process; Not tested in realistic scenarios; Poor system's coverage; Tested only with UAV helicopters 需要电缆和校准过程；未在现实场景中进行测试；系统覆盖范围小；仅使用无人驾驶飞行器直升机进行过测试
Voice/Data communications; Insertion of important information (casualty; valuable equipment; or dangerous material); Integration with external alert systems; Monitoring environmental parameters; Records the environment 语音/数据通信；插入重要信息（伤员、贵重设备或危险材料）；与外部警报系统整合；监测环境参数；记录环境情况
Li et al. - 2014 [77] 李等人 - 2014 [77]
Yes 是
Limited tests; The low availability of BIM; Reliability of the smartphone communications with the web server; Time needed to process the information of BIM models; The framework cannot be used with inertial navigation systems. 测试有限；BIM 可用性低；智能手机与网络服务器通信的可靠性；处理 BIM 模型信息所需的时间；该框架不能与惯性导航系统一起使用。
The framework can be used with other RF technologies; The BIM tool provide a GUI 该框架可与其他射频技术一起使用；BIM 工具提供图形用户界面
Outside communication; COTS components; Positioning information embedded on the SSID; Track ~~800\approx 800 devices simultaneously; Position update of 10 s ; Low computational complexity 外部通信；COTS 组件；SSID 上嵌入定位信息；同时跟踪 ~~800\approx 800 设备；10 秒内更新位置；计算复杂度低
IMU-Based Systems 基于 IMU 的系统
Beauregard - 2006 [80] 博雷加德 - 2006 [80]
No 没有
Specific calibration process; IMU must be mounted in a fixed position; Heading errors; The system cannot differentiate the gaze orientation from the direction of travel 特定的校准过程；IMU 必须安装在固定位置；方向误差；系统无法区分注视方向和行进方向
Helmet-mounted IMU 头盔式 IMU
Beauregard - 2007 [81] 博雷加德 - 2007 [81]
No 没有
PDR drift; Limited tests PDR 漂移；有限测试
Identifies the typical firefighters' movements 识别消防员的典型动作
IndoorNav - 2007 [82]
Yes 是
PDR drift; Error associated to the tag attached to the door; Door's omission; Tags can be destroyed by the fire PDR 漂移；与门上的标签相关的错误；门的遗漏；标签可能被火烧毁
Posture's monitoring 姿势监测
Ojeda and Borenstein 2007 [83] Ojeda 和 Borenstein 2007 [83]
No 没有
Heading errors grows over time; The IMU is too large to fit in the sole of a boot; Sometimes it detects false footfalls 随着时间的推移，方向误差会逐渐增大；IMU 体积太大，无法装入靴底；有时会检测到错误的脚步声
Independent of any infrastructure; Zero-radiation signature 独立于任何基础设施；零辐射签名
Widyawan et al. -2008 [84] Widyawan 等人-2008 [84]
Yes 是
PDR drift PDR 漂移
- ^("cosen "){ }^{\text {cosen }}
HeadSLAM - 2008 [86]
No 没有
PDR drift; The laser scanner fails in low visibility environments PDR 漂移；激光扫描仪在能见度低的环境中失灵
Monitoring environmental parameters; Communication; Construction of the environment 监测环境参数；交流；环境建设
CADMS - 2010 [89]
Yes 是
The low availability of BIM; Effort to keep BIM up-to-date BIM 可用性低；努力更新 BIM
GUI; Communication; Independent of any infrastructure 图形用户界面；通信；独立于任何基础设施
PDR drift; Does not have a support for the communication with incident commander PDR 漂移；不支持与事故指挥官的通信
Tracking the body movement; Dynamic threshold selection for the ZUPT method according to the type of movement; Works with both walking and running motions 跟踪身体运动；根据运动类型动态选择 ZUPT 方法的阈值；适用于步行和跑步运动
Hari el al. - 2013 [94] Hari el al.- 2013 [94]
Optional 可选
Requires a local infrastructure Wi-Fi network to send the processed data to the incident commander; Needs inter-agent ranging 需要本地基础设施 Wi-Fi 网络将处理过的数据发送给事件指挥员；需要各机构之间进行测距
Monitoring firefighter's intervention through a helmet-mounted camera 通过头盔上的摄像头监控消防员的干预行为
REFIRE - 2013 [13] 重燃 - 2013 [13]
No 没有
Requires the pre-installation and location mapping of the RFID tags; Relies on user's specific parameters; The step detection based on time windows techniques are prone to fail in face of irregular movements. Tags can be destroyed by the fire 需要预先安装 RFID 标签并绘制位置图；依赖于用户的特定参数；基于时间窗口的步骤检测技术在面对不规则移动时容易失效。标签可能被火烧毁
The RFID tags can store critical up-to-date building information RFID 标签可存储重要的最新楼宇信息
Studied different types of firefighter's movements during a typical intervention; Professional firefighters were used to evaluate the system's performance 研究消防员在典型干预过程中的不同动作类型；使用专业消防员评估系统性能
TOR-2014 [100]
No 没有
Requires a local infrastructure Wi-Fi network to send the processed data to the incident commander; Needs inter-agent ranging; Needs a calibration process 需要本地基础设施 Wi-Fi 网络将处理过的数据发送给事件指挥员；需要机构间测距；需要校准过程
Evaluated under different movements typically of firefighters; Professional firefighters were used to evaluate the system 在消防员典型的不同动作下进行评估；使用专业消防员对系统进行评估
The beacons can be moved or destroyed; PDR drift 信标可以移动或销毁；PDR 漂移
Identifies the typical firefighter's movements 识别消防员的典型动作
GLANSER - 2012 [109] 格兰瑟 - 2012 [109]
Optional 可选
System under development 系统正在开发中
Outside communication; Identifies the typical firefighters' movements; Data processing implemented in FPGA. Monitoring physiological parameters 外部通信；识别消防员的典型动作；通过 FPGA 实现数据处理。监测生理参数
WASP - 2013 [112] 黄蜂 - 2013 [112]
Yes 是
-
Construction of the environment; Patented sensor fusion & map building technology for robust, accurate GPS-denied location 构建环境；获得专利的传感器融合和地图构建技术，可在 GPS 不存在的情况下进行稳健、准确的定位
Simon el al. 2015 [114] Simon el al.2015 [114]
No 没有
The handheld device is required to stand still during the position calculation; The handheld device interferes with firefighters' mission; The landmarks are vulnerable to high temperatures; Requires landmarks preinstalled on the building 计算位置时要求手持设备静止不动；手持设备干扰消防员执行任务；地标易受高温影响；要求在建筑物上预装地标
The landmarks have a long operating time ( ~~8\approx 8 years); The system is fully operational even in catastrophic scenarios 地标运行时间长（ ~~8\approx 8 年）；即使在灾难性情况下，系统也能完全正常运行
Other Systems 其他系统
PathFinder [116] 路径查找器 [116］
No 没有
The sound can be corrupted; The direction is difficult to determine 声音可能被破坏；方向难以确定
Only work for slow movements; The Smart Lifeline can be trapped on the building 仅适用于缓慢移动；智能生命线可能会被困在建筑物上
Monitoring of physiological and environmental parameters; Outside communication 监测生理和环境参数；外部通信
System/Author Name Year Layout Limitations Other Features
Radio Signal-Based Systems
SmokeNet-2007 [65] Yes Sensitive to changes in the environment; Requires calibration Monitoring environmental parameters; Outside communication.
" Europcom - 2008 [17],
[66], [67] " Optional The accuracy is not always guaranteed Outside communication
ProeTEX - 2009 [68] No Low performance in indoor Monitoring physiological and environmental parameters
LifeNet-2009 [70] No The beacons can be moved or destroyed; The position's information is relative. Short range communication (2,5m) Outside communication. Monitoring physiological and environmental parameters. Assisted navigation
FIREGUIDE - 2010 [71] Yes Sensitive to changes in the environment; Depends on building power Indication of the nearest exit; Outside communication
Ruppel el al. - 2010 [72] Yes Limited tests; The low availability of BIM models; Expensive site survey; Relies on building power Records the firefighters' operation
Li and Gerber - 2012 [20] No Room-scale tested with LOS conditions; Requires absolute position; The propagation model parameters are site specific; The App needs to be installed on the smartphone's occupants Localization of both first responders and occupants
Moon et al. - 2013 [74], [75] Yes Very limited tests; The building structured can be outdated Outside communication; Initialization mechanism
Ghosh et al. - 2013 [76] No System under development; Not evaluated Monitoring of physiological and environmental parameters; Prediction of the fire expansion
Zhang et al. - 2013 [53] Optional Requires cables and a calibration process; Not tested in realistic scenarios; Poor system's coverage; Tested only with UAV helicopters Voice/Data communications; Insertion of important information (casualty; valuable equipment; or dangerous material); Integration with external alert systems; Monitoring environmental parameters; Records the environment
Li et al. - 2014 [77] Yes Limited tests; The low availability of BIM; Reliability of the smartphone communications with the web server; Time needed to process the information of BIM models; The framework cannot be used with inertial navigation systems. The framework can be used with other RF technologies; The BIM tool provide a GUI
Femminella and Reali - 2015" [18] " Optional Low accuracy Outside communication; COTS components; Positioning information embedded on the SSID; Track ~~800 devices simultaneously; Position update of 10 s ; Low computational complexity
IMU-Based Systems
Beauregard - 2006 [80] No Specific calibration process; IMU must be mounted in a fixed position; Heading errors; The system cannot differentiate the gaze orientation from the direction of travel Helmet-mounted IMU
Beauregard - 2007 [81] No PDR drift; Limited tests Identifies the typical firefighters' movements
IndoorNav - 2007 [82] Yes PDR drift; Error associated to the tag attached to the door; Door's omission; Tags can be destroyed by the fire Posture's monitoring
Ojeda and Borenstein 2007 [83] No Heading errors grows over time; The IMU is too large to fit in the sole of a boot; Sometimes it detects false footfalls Independent of any infrastructure; Zero-radiation signature
Widyawan et al. -2008 [84] Yes PDR drift - ^("cosen ")
HeadSLAM - 2008 [86] No PDR drift; The laser scanner fails in low visibility environments Monitoring environmental parameters; Communication; Construction of the environment
CADMS - 2010 [89] Yes The low availability of BIM; Effort to keep BIM up-to-date GUI; Communication; Independent of any infrastructure
Rantakokko et al. - 2011 [90] No Limited tests; Needs inter-agent ranging -
Zhang et al. - 2012 [92] No PDR drift; Does not have a support for the communication with incident commander Tracking the body movement; Dynamic threshold selection for the ZUPT method according to the type of movement; Works with both walking and running motions
Hari el al. - 2013 [94] Optional Requires a local infrastructure Wi-Fi network to send the processed data to the incident commander; Needs inter-agent ranging Monitoring firefighter's intervention through a helmet-mounted camera
REFIRE - 2013 [13] No Requires the pre-installation and location mapping of the RFID tags; Relies on user's specific parameters; The step detection based on time windows techniques are prone to fail in face of irregular movements. Tags can be destroyed by the fire The RFID tags can store critical up-to-date building information
Rantakokko et al. - 2014 [19] No Tested only in straight line movement Studied different types of firefighter's movements during a typical intervention; Professional firefighters were used to evaluate the system's performance
TOR-2014 [100] No Requires a local infrastructure Wi-Fi network to send the processed data to the incident commander; Needs inter-agent ranging; Needs a calibration process Evaluated under different movements typically of firefighters; Professional firefighters were used to evaluate the system
Hybrid Systems
"Relate Trails - 2008
, [101], [102]" No Limited tests; The beacons can be moved or destroyed; PDR drift Monitoring environmental parameters; Outside Communication
PPL-2011 [103]-[107] No System tested on a wheeled platform; Does not detect footsteps The localization algorithm is implemented in FPGA
" Virtual Lifeline - 2011
[108]" Optional The beacons can be moved or destroyed; PDR drift Identifies the typical firefighter's movements
GLANSER - 2012 [109] Optional System under development Outside communication; Identifies the typical firefighters' movements; Data processing implemented in FPGA. Monitoring physiological parameters
WASP - 2013 [112] Yes - Construction of the environment; Patented sensor fusion & map building technology for robust, accurate GPS-denied location
Simon el al. 2015 [114] No The handheld device is required to stand still during the position calculation; The handheld device interferes with firefighters' mission; The landmarks are vulnerable to high temperatures; Requires landmarks preinstalled on the building The landmarks have a long operating time ( ~~8 years); The system is fully operational even in catastrophic scenarios
Other Systems
PathFinder [116] No The sound can be corrupted; The direction is difficult to determine -
MILPS [117]-[120] No Low range -
Smart Lifeline - 2013 [121], [122] No Only work for slow movements; The Smart Lifeline can be trapped on the building Monitoring of physiological and environmental parameters; Outside communication| System/Author Name Year | Layout | Limitations | Other Features |
| :---: | :---: | :---: | :---: |
| Radio Signal-Based Systems | | | |
| SmokeNet-2007 [65] | Yes | Sensitive to changes in the environment; Requires calibration | Monitoring environmental parameters; Outside communication. |
| $\begin{aligned} & \text { Europcom - } 2008 \text { [17], } \\ & \text { [66], [67] } \end{aligned}$ | Optional | The accuracy is not always guaranteed | Outside communication |
| ProeTEX - 2009 [68] | No | Low performance in indoor | Monitoring physiological and environmental parameters |
| LifeNet-2009 [70] | No | The beacons can be moved or destroyed; The position's information is relative. Short range communication $(2,5 \mathrm{~m})$ | Outside communication. Monitoring physiological and environmental parameters. Assisted navigation |
| FIREGUIDE - 2010 [71] | Yes | Sensitive to changes in the environment; Depends on building power | Indication of the nearest exit; Outside communication |
| Ruppel el al. - 2010 [72] | Yes | Limited tests; The low availability of BIM models; Expensive site survey; Relies on building power | Records the firefighters' operation |
| Li and Gerber - 2012 $[20]$ | No | Room-scale tested with LOS conditions; Requires absolute position; The propagation model parameters are site specific; The App needs to be installed on the smartphone's occupants | Localization of both first responders and occupants |
| Moon et al. - 2013 [74], [75] | Yes | Very limited tests; The building structured can be outdated | Outside communication; Initialization mechanism |
| Ghosh et al. - 2013 [76] | No | System under development; Not evaluated | Monitoring of physiological and environmental parameters; Prediction of the fire expansion |
| Zhang et al. - 2013 [53] | Optional | Requires cables and a calibration process; Not tested in realistic scenarios; Poor system's coverage; Tested only with UAV helicopters | Voice/Data communications; Insertion of important information (casualty; valuable equipment; or dangerous material); Integration with external alert systems; Monitoring environmental parameters; Records the environment |
| Li et al. - 2014 [77] | Yes | Limited tests; The low availability of BIM; Reliability of the smartphone communications with the web server; Time needed to process the information of BIM models; The framework cannot be used with inertial navigation systems. | The framework can be used with other RF technologies; The BIM tool provide a GUI |
| Femminella and Reali - $2015 \text { [18] }$ | Optional | Low accuracy | Outside communication; COTS components; Positioning information embedded on the SSID; Track $\approx 800$ devices simultaneously; Position update of 10 s ; Low computational complexity |
| IMU-Based Systems | | | |
| Beauregard - 2006 [80] | No | Specific calibration process; IMU must be mounted in a fixed position; Heading errors; The system cannot differentiate the gaze orientation from the direction of travel | Helmet-mounted IMU |
| Beauregard - 2007 [81] | No | PDR drift; Limited tests | Identifies the typical firefighters' movements |
| IndoorNav - 2007 [82] | Yes | PDR drift; Error associated to the tag attached to the door; Door's omission; Tags can be destroyed by the fire | Posture's monitoring |
| Ojeda and Borenstein 2007 [83] | No | Heading errors grows over time; The IMU is too large to fit in the sole of a boot; Sometimes it detects false footfalls | Independent of any infrastructure; Zero-radiation signature |
| Widyawan et al. -2008 [84] | Yes | PDR drift | - ${ }^{\text {cosen }}$ |
| HeadSLAM - 2008 [86] | No | PDR drift; The laser scanner fails in low visibility environments | Monitoring environmental parameters; Communication; Construction of the environment |
| CADMS - 2010 [89] | Yes | The low availability of BIM; Effort to keep BIM up-to-date | GUI; Communication; Independent of any infrastructure |
| Rantakokko et al. - 2011 [90] | No | Limited tests; Needs inter-agent ranging | - |
| Zhang et al. - 2012 [92] | No | PDR drift; Does not have a support for the communication with incident commander | Tracking the body movement; Dynamic threshold selection for the ZUPT method according to the type of movement; Works with both walking and running motions |
| Hari el al. - 2013 [94] | Optional | Requires a local infrastructure Wi-Fi network to send the processed data to the incident commander; Needs inter-agent ranging | Monitoring firefighter's intervention through a helmet-mounted camera |
| REFIRE - 2013 [13] | No | Requires the pre-installation and location mapping of the RFID tags; Relies on user's specific parameters; The step detection based on time windows techniques are prone to fail in face of irregular movements. Tags can be destroyed by the fire | The RFID tags can store critical up-to-date building information |
| Rantakokko et al. - 2014 [19] | No | Tested only in straight line movement | Studied different types of firefighter's movements during a typical intervention; Professional firefighters were used to evaluate the system's performance |
| TOR-2014 [100] | No | Requires a local infrastructure Wi-Fi network to send the processed data to the incident commander; Needs inter-agent ranging; Needs a calibration process | Evaluated under different movements typically of firefighters; Professional firefighters were used to evaluate the system |
| Hybrid Systems | | | |
| Relate Trails - 2008 <br> , [101], [102] | No | Limited tests; The beacons can be moved or destroyed; PDR drift | Monitoring environmental parameters; Outside Communication |
| PPL-2011 [103]-[107] | No | System tested on a wheeled platform; Does not detect footsteps | The localization algorithm is implemented in FPGA |
| $\begin{aligned} & \text { Virtual Lifeline - } 2011 \\ & {[108]} \end{aligned}$ | Optional | The beacons can be moved or destroyed; PDR drift | Identifies the typical firefighter's movements |
| GLANSER - 2012 [109] | Optional | System under development | Outside communication; Identifies the typical firefighters' movements; Data processing implemented in FPGA. Monitoring physiological parameters |
| WASP - 2013 [112] | Yes | - | Construction of the environment; Patented sensor fusion & map building technology for robust, accurate GPS-denied location |
| Simon el al. 2015 [114] | No | The handheld device is required to stand still during the position calculation; The handheld device interferes with firefighters' mission; The landmarks are vulnerable to high temperatures; Requires landmarks preinstalled on the building | The landmarks have a long operating time ( $\approx 8$ years); The system is fully operational even in catastrophic scenarios |
| Other Systems | | | |
| PathFinder [116] | No | The sound can be corrupted; The direction is difficult to determine | - |
| MILPS [117]-[120] | No | Low range | - |
| Smart Lifeline - 2013 [121], [122] | No | Only work for slow movements; The Smart Lifeline can be trapped on the building | Monitoring of physiological and environmental parameters; Outside communication |
localization of emergency responders during their missions. This conclusion is supported on the systematic and in-depth study conducted on this survey paper on IPSs for emergency 应急响应人员在执行任务期间的定位。本调查报告对用于紧急情况的 IPS 系统进行了系统和深入的研究，并得出了上述结论。
responders. Since the best performing IPSs developed for emergency responders are based on proprietary technologies, specially designed for the system, the cost of the IPS is high, 应急人员。由于为应急响应人员开发的性能最好的 IPS 是基于专有技术、专门为该系统设计的，因此 IPS 的成本很高、
therefore, unaffordable to the majority of emergency responders. As analyzed in Section III-F, the cost of an IPS is a major issue for its acceptance and an important factor that researchers/designers of IPS for emergency responders have to consider when selecting the technology or combination of technologies to be used. 因此，大多数应急响应人员负担不起。正如第 III-F 节所分析的，IPS 的成本是影响其接受程度的一个主要问题，也是应急响应者 IPS 研究人员/设计人员在选择使用的技术或技术组合时必须考虑的一个重要因素。
2) Test Scenarios: The majority of the IPSs are evaluated under controlled or semi-controlled environments with predefined routes or movements. Therefore, the performance of an IPS is tailored for a single testbed environment and for specific evaluation criteria [15], [123]. 2) 测试场景：大多数 IPS 都是在受控或半受控环境下进行评估的，并预先设定了路线或移动方式。因此，IPS 的性能是针对单一测试平台环境和特定评估标准量身定制的 [15]，[123]。
In real scenarios, a huge number of unpredictable situations may occur which can lead to a failure on the IPS. These unpredictable situations include the destruction of waypoints or landmarks by the flames, unrecognizable movement pattern, building geometry, coverage area, among others. So, for a correct assessment of the IPS performance, it has to be evaluated under real scenarios and according to evaluation criteria previously defined by emergency responders. Only in this way the required performance of the IPS is achievable and the limitations of the existing IPSs for emergency responders can be overcome. Additionally, the comparison of several IPSs performances under the same conditions will be extremely helpful to study the effects that environmental conditions (temperature and humidity) and building structures have in the technology selected for each IPS [123]. 在实际场景中，可能会出现大量不可预测的情况，导致 IPS 出现故障。这些不可预测的情况包括航点或地标被火焰摧毁、运动模式无法识别、建筑物几何形状、覆盖区域等。因此，为了正确评估 IPS 的性能，必须在真实场景下并根据应急响应人员事先确定的评估标准对其进行评估。只有这样，IPS 才能达到所要求的性能，才能克服现有 IPS 在应急响应方面的局限性。此外，在相同条件下比较几种 IPS 的性能，对于研究环境条件（温度和湿度）和建筑结构对每种 IPS 所选技术的影响非常有帮助[123]。
3) Recognition of Typical Movement Patterns of Emergency Responders: As stated by Harris, there is a gap between what researchers are developing and what emergency responders expect from an IPS [25]. That happens due to the lack of cooperation between the IPS developers and the emergency responders, resulting in IPSs that achieve good results in controlled environments but, in real scenarios, are incapable of producing good position estimations. 3) 识别应急响应人员的典型运动模式：正如 Harris 所说，研究人员开发的 IPS 与应急响应人员对 IPS 的期望之间存在差距[25]。这是因为 IPS 开发人员与应急响应人员之间缺乏合作，导致 IPS 在受控环境中取得了良好的效果，但在真实场景中却无法产生良好的位置估计。
In addition to the unique challenges created by an environment where the missions of emergency responders take place (e.g., unstructured, poor visibility, high temperature and humidity), another differentiating characteristic between normal IPSs and IPSs for emergency responders is the unique movement patterns of the emergency responders during their missions. However, most of the developed IPSs are incapable of recognizing these characteristic movement patterns e.g., walk (upright or hunched), “knee-dragging motion”, or crawling [19]- consequently, of producing good position estimations with this type of movements. As pointed out by Harris, this is the main cause of the IPSs failure when tested under typical emergency responders’ missions [25]. Usually, the majority of the developed IPSs for emergency responders were tested on wheeled platforms or for simple movements (e.g., walking, running and climbing stairs). 除了应急响应人员执行任务的环境（如非结构化、能见度低、高温高湿）所带来的独特挑战外，普通 IPS 与应急响应人员 IPS 之间的另一个区别特征是应急响应人员在执行任务期间的独特运动模式。然而，大多数已开发的 IPS 无法识别这些特有的运动模式，如行走（直立或蜷缩）、"拖膝运动 "或爬行[19]，因此，也就无法对这类运动进行良好的位置估计。正如 Harris 所指出的，这是 IPS 在典型的应急响应任务下测试失败的主要原因[25]。通常，大多数为应急响应人员开发的 IPS 都是在轮式平台或简单运动（如步行、跑步和爬楼梯）中进行测试的。
4) Poor Availability of Signals From Wireless Technologies: In indoor scenarios, radio signals are subjected to countless physical phenomena - e.g., multipath, obstruction, reflection, scattering and diffraction - which drastically attenuate the signal, reducing its coverage range. So, to have a good availability of radio signal throughout the intervention scenario, several anchor nodes have to be distributed over the place. The anchor nodes density will be as higher as lower the wireless 4) 无线技术信号的可用性差：在室内场景中，无线电信号会受到无数物理现象的影响，如多径、阻塞、反射、散射和衍射等，这些物理现象会极大地衰减信号，从而缩小信号的覆盖范围。因此，要想在整个干预场景中提供良好的无线电信号，就必须在各地分布多个锚节点。锚节点的密度越高，无线网络的覆盖范围就越小。
technology range is. Multiple anchor nodes are needed because most of the localization methods based on geometric properties or fingerprinting require at least three different signals for the localization process. However, fulfilling this requirement in an unstructured environment is almost impossible, as the anchor nodes, if used, must be deployed at the moment of intervention and their deployment should not interfere with on-duty tasks of emergency responders [17]-[20], [22], [28]. 技术范围。之所以需要多个锚节点，是因为大多数基于几何特性或指纹识别的定位方法在定位过程中至少需要三个不同的信号。然而，要在非结构化环境中满足这一要求几乎是不可能的，因为如果使用锚节点，则必须在干预时部署，而且其部署不应干扰应急响应人员的值班任务 [17]-[20]、[22]、[28]。
An alternative to the use of wireless signals are the inertial sensors. However, pure IMU-based systems only provide an accurate position estimate for short periods. For long periods of operation, which is a requirement of emergency responders (Section III-E), the error in the position estimation will grow exponentially, making this information useless. In IMU-based systems, the error in position estimation is associated with the sensors’ drift. For long-term operations, the position estimation obtained from IMU data must be periodically corrected with the information gathered from wireless technologies, waypoints, landmarks, or maps. 惯性传感器是无线信号的替代品。然而，纯粹基于 IMU 的系统只能在短时间内提供准确的位置估计。对于应急响应人员所要求的长时间操作（第 III-E 节），位置估算的误差将呈指数增长，从而使该信息毫无用处。在基于 IMU 的系统中，位置估计误差与传感器的漂移有关。为了实现长期运行，必须定期使用从无线技术、航点、地标或地图中收集的信息来校正从 IMU 数据中获得的位置估计值。
B. Future Research Directions B.未来研究方向
A detailed survey of typical techniques, methods and approaches used in IPSs for emergency responders has been presented in this study. However, many design aspects require investigation to increase the performance and usability of IPSs for emergency responders. In this section, some key future research directions in IPSs for emergency responders are discussed. 本研究详细介绍了用于应急响应人员的 IPS 的典型技术、方法和途径。然而，要提高应急响应人员 IPS 的性能和可用性，还需要对许多设计方面进行研究。本节将讨论面向应急响应人员的 IPS 的一些关键未来研究方向。
Distribution Criteria of Waypoints: The use of waypoints is a good solution to overcome the poor availability of signals from wireless technologies and the drift of inertial sensors. Therefore, waypoints are a cost-effective solution to increase the performance of all three IPS variants discussed in the taxonomy proposed in this survey paper-radio signal-based, IMU-based and hybrid systems. Additionally, waypoints can also be used to increase the coverage of an IPS. 航点分布标准：使用航点是克服无线技术信号可用性差和惯性传感器漂移的好办法。因此，航点是一种具有成本效益的解决方案，可提高本调查报告提出的分类法中讨论的所有三种 IPS 变体（基于无线电信号的系统、基于 IMU 的系统和混合系统）的性能。此外，航点还可用于扩大 IPS 的覆盖范围。
Nevertheless, waypoints have to be deployed as an emergency responder enters a building and its position must be computed on-the-fly. The overall performance of the IPS will be intrinsically dependent on the computed position accuracy of the waypoint. Knowing when a waypoint should be deployed is a crucial factor in the performance of the IPS. On the one hand, if the waypoints are deployed sparsely, the position error may be too high to meet the requirements in terms of localization accuracy (Section III-A). On the other hand, if the waypoints are deployed too often, the deployment process will interfere with the on-duty tasks of emergency responders, which is undesirable (Section III-C). So, further research on algorithms for positioning and assisted deployment of waypoints is needed. Alongside with the development of these algorithms, the researchers should also focus on the selection of the waypoint wireless technology. This technology should be selected to satisfy a trade-off between reducing the number of waypoints to be deployed and maintaining the localization error below the specified requirement. The assisted deployment algorithms can be developed monitoring the position 然而，当应急响应人员进入建筑物时，必须部署航点，并且必须即时计算航点的位置。IPS 的整体性能本质上取决于计算出的航点位置精度。了解何时应部署航点是影响 IPS 性能的关键因素。一方面，如果航点部署稀疏，位置误差可能过大，无法满足定位精度的要求（第 III-A 节）。另一方面，如果航点部署过于频繁，部署过程将干扰应急响应人员的执勤任务，这也是不可取的（第 III-C 节）。因此，需要进一步研究航点定位和辅助部署算法。在开发这些算法的同时，研究人员还应关注航点无线技术的选择。在选择该技术时，应在减少要部署的航点数量和保持定位误差低于规定要求之间进行权衡。辅助部署算法的开发可通过对位置进行监测来实现。
error - i.e., a waypoint is deployed every time the position error exceeds a specified threshold - or when the RSS from another waypoint drops below a specified threshold. 即每当位置误差超过指定阈值时，或当另一个航点的 RSS 下降到指定阈值以下时，就部署一个航点。
2) Development of a Localization Algorithm: Albeit several IPSs have been proposed to support emergency responders during their daily missions, none of them is capable of providing a satisfactory performance in all environments [14]-[16]. This happens mainly due to the limitations of the existing technologies - both, the wireless technologies and the inertial sensors -, the challenging environmental conditions and the diversity and complexity of building structures. As discussed in the Section III-F, the development of new technologies to tackle the indoor localization problem is not a suitable approach as it will increase the cost of the system, making the IPS too expensive for the emergency responders. 2) 开发定位算法：尽管已经提出了几种惯性传感器来支持应急响应人员的日常任务，但没有一种惯性传感器能够在所有环境中提供令人满意的性能[14]-[16]。这主要是由于现有技术（无线技术和惯性传感器）的局限性、具有挑战性的环境条件以及建筑结构的多样性和复杂性造成的。正如第 III-F 部分所讨论的，开发新技术来解决室内定位问题并不是一种合适的方法，因为这会增加系统的成本，使 IPS 对于应急响应人员来说过于昂贵。
So, a cost-effective solution is the combination of different technologies, techniques and methods with data fusion algorithms [48], [80]. These technologies, techniques and methods should be chosen in order to complement the drawbacks of each. For instance, inertial sensors can be used to overcome the poor availability of wireless signals in indoor scenarios, therefore, providing a continuous position estimation; waypoints can be used to control the drift of inertial sensors; or the CSI can be used to replace the traditional RSS, as the CSI-based techniques outperform the RSSI-based techniques without having to change the localization method of the IPS [59]-[61]. The combination of different technologies (e.g., GPS with inertial sensors and Wi-Fi) allows the localization of emergency responders in both indoor and outdoor scenarios. Additionally, different sensors can also be used to decrease the complexity of the localization process, e.g., barometers or atmospheric pressure sensors can be used to provide the vertical position. 因此，一种具有成本效益的解决方案是将不同的技术、工艺和方法与数据融合算法相结合 [48]，[80]。在选择这些技术、工艺和方法时，应注意补充每种技术、工艺和方法的缺点。例如，可以利用惯性传感器克服室内场景中无线信号可用性差的问题，从而提供连续的位置估计；可以利用航点控制惯性传感器的漂移；也可以利用 CSI 取代传统的 RSS，因为基于 CSI 的技术优于基于 RSSI 的技术，而无需改变 IPS 的定位方法[59]-[61]。不同技术（如 GPS 与惯性传感器和 Wi-Fi）的结合可在室内和室外场景中对应急响应人员进行定位。此外，还可以使用不同的传感器来降低定位过程的复杂性，例如，可以使用气压计或大气压力传感器来提供垂直位置。
However, data fusion algorithms require an error model that combines all the error sources that might affect the IPS performance. By itself, modeling the system error based on the data provided by the several sensors that comprise the IPS is an attractive research topic that needs further research [53]. Therefore, the data fusion algorithm should adequately combine the information provided by the different sensors with the probabilistic integration of these data over the time. 然而，数据融合算法需要一个误差模型，将可能影响 IPS 性能的所有误差源结合起来。根据组成 IPS 的多个传感器提供的数据建立系统误差模型本身就是一个有吸引力的研究课题，需要进一步研究[53]。因此，数据融合算法应充分结合不同传感器提供的信息以及这些数据随时间变化的概率整合。
3) Creation of Benchmarks and Test Scenarios: As discussed in the Section VII-A, one of the biggest challenges in the performance comparison of different IPSs is the lack of benchmarks. Besides providing a correct assessment of the IPSs’ performances, benchmarks are also regarded as an essential tool for the identification of the real limitations of technologies, techniques and methods when applied in indoor environments [15], [16], [25], [100], [123], [124]. 3) 创建基准和测试场景：如第 VII-A 节所述，比较不同 IPS 性能的最大挑战之一是缺乏基准。除了对 IPS 的性能进行正确评估外，基准还被认为是确定技术、工艺和方法在室内环境中应用的实际局限性的重要工具[15], [16], [25], [100], [123], [124]。
Although some researchers proposed platforms to benchmark the IPSs [123], [124] and international competitions for performance assessment of IPSs were created [15], [16], benchmarks are still in an initial phase and do not address the challenging scenarios faced by IPSs designed for emergency responders. Alongside with the need of benchmarks for the comparison on different IPSs, it is also required to create standard test scenarios to portray the conditions that the emergency responders face during their missions [25]. 尽管一些研究人员提出了制定 IPS 基准的平台[123]、[124]，并创建了 IPS 性能评估国际竞赛[15]、[16]，但基准仍处于初始阶段，并没有解决为应急响应人员设计的 IPS 所面临的挑战性情景。除了需要基准来比较不同的 IPS 外，还需要创建标准的测试场景来描绘应急响应人员在执行任务时所面临的条件[25]。
4) Integration of an IPS in a Cyber-Physical System: The integration of several different technologies in one system led to a new type of systems, the Cyber-Physical Systems (CPS). A CPS consists of a distributed intelligent system, composed of embedded systems - sensors and actuators designed to sense and interact with the physical world, including human users - in order to collect data globally, process the information centrally, and distribute the results locally [125], [126]. A CPS aims to enhance the situational awareness and the safety of the user. 4) 将 IPS 集成到网络物理系统中：将几种不同的技术集成到一个系统中，产生了一种新型系统，即网络物理系统（CPS）。CPS 由分布式智能系统组成，由嵌入式系统--传感器和执行器--组成，旨在感知物理世界（包括人类用户）并与之互动，以便在全球范围内收集数据，集中处理信息，并在本地发布结果 [125]，[126]。CPS 旨在提高用户的态势感知能力和安全性。
The development of an IPS integrated on a CPS, which gathers information from many other environmental and physiological sensors, has many advantages, namely: 1) refinement of the emergency responder’s position based on the positions of the other emergency responders; 2) creation of new escape routes for each emergency responder, based on the received environmental information and position estimation; 3) reduction of the computational capacities through the distribution of the computational workload; 4) a system capable of interpreting and adjusting in real-time to the dynamic changes of the environment, and 5) increasing the situational awareness of emergency responders and the incident commander, providing not only the position information but also the physiological and environmental conditions. 集成在 CPS 上的 IPS 可收集来自许多其他环境和生理传感器的信息，其开发具有许多 优势，即1) 根据其他应急响应者的位置，完善应急响应者的位置；2) 根据接收到的环境信息和位置估计，为每个应急响应者创建新的逃生路线；3) 通过分配计算工作量，降低计算能力；4) 系统能够实时解释和调整环境的动态变化；5) 不仅提供位置信息，还提供生理和环境条件，提高应急响应者和事故指挥官的态势感知能力。
Despite the advantages of the integration of an IPS in a CPS, until now only the Smart Firefighting Project proposed an integrated system with the localization feature [125]. However, so far, the findings were not published. 尽管将 IPS 集成到 CPS 中具有诸多优势，但到目前为止，只有智能消防项目提出了一个具有定位功能的集成系统[125]。但迄今为止，相关研究结果尚未公布。
5) Integration of Existing Data in the IPS: A recent trend in the indoor localization field is the use of crowdsourcing strategies to construct user trajectories based on the data gathered from mobile devices equipped with sensors [127]-[129]. Based on the user trajectories, an indoor map can be created. 5) 在 IPS 中整合现有数据：室内定位领域的一个最新趋势是使用众包策略，根据从配备传感器的移动设备收集的数据构建用户轨迹 [127]-[129]。根据用户轨迹，可以创建室内地图。
As pointed out several times throughout this survey paper, an IPS for emergency responders cannot be based on preinstalled systems, detailed building maps, or even pre-collected information [17]-[19]. Nevertheless, the information on the user trajectories inside a building can be extremely helpful to increase the performance of an IPS. Based on the user trajectories some useful information can be extracted: 1) identification of building structures (e.g., corridors, canteen, elevators and stairs), which are useful for planning the intervention and for the generation of escape routes; 2) identification of the places with the highest likelihood of having occupants. Setting these places with the highest priority will reduce the time needed to search occupants; and 3) creation of digital building maps which may provide an initial estimation of the building dimensions and navigation functionalities. These maps may be refined as the emergency responders enter the building. 正如本文多次指出的那样，为应急响应人员设计的 IPS 不能基于预安装的系统、详细的建筑地图，甚至是预先收集的信息 [17]-[19]。然而，建筑物内的用户轨迹信息对提高 IPS 性能非常有帮助。根据用户轨迹可以提取一些有用的信息：1）识别建筑结构（如走廊、食堂、电梯和楼梯），这对规划干预措施和生成逃生路线非常有用；2）识别最有可能有用户的地方。将这些地方设定为最优先位置将减少搜索人员所需的时间；以及 3) 绘制数字楼宇地图，以便对楼宇尺寸和导航功能进行初步估算。这些地图可在应急响应人员进入大楼后进行完善。
To achieve the goal, more research is needed addressing the following issues: 1) how to extract user trajectories and building structures from the crowdsourcing data; 2) how to deal with the errors generated from the low-cost sensors that equip most of the mobile devices like smartphones; 3) which information is useful and can be later used in digital building maps for emergency responders during their missions; and 4) how to combine the created digital maps and the user trajectories with the IPS. 为实现这一目标，需要针对以下问题开展更多研究：1) 如何从众包数据中提取用户轨迹和建筑结构；2) 如何处理大多数移动设备（如智能手机）所配备的低成本传感器产生的误差；3) 哪些信息是有用的，可以在数字建筑地图中使用，以便应急响应人员在执行任务时使用；以及 4) 如何将创建的数字地图和用户轨迹与 IPS 结合起来。
VIII. Conclusion VIII.结束语
Throughout this work, the demand for an accurate, reliable, scalable, affordable and autonomous IPS for emergency responders was highlighted. The information provided by an IPS that is capable of, simultaneous, positioning and tracking emergency responders on the field is of utmost importance as it allows: 1) increasing the situational awareness of both, the emergency responders and the incident commander; 2) supporting the decision-making process and the mission planning; and 3) generating escape routes and assisted navigation. Therefore, an IPS is regarded by the emergency responders as an essential tool to decrease the number of deaths during on-duty missions and to make the planning and on-site coordination more efficient. 在整个工作过程中，强调了应急响应人员对准确、可靠、可扩展、经济实惠和自主的 IPS 的需求。能够同时定位和跟踪现场应急响应人员的 IPS 所提供的信息至关重要，因为它能够：1）提高应急响应人员和事件指挥官对态势的认识；2）提高应急响应人员和事件指挥官对态势的认识；3）提高应急响应人员和事件指挥官对态势的认识：1) 提高应急人员和事故指挥官对态势的认识；2) 支持决策过程和任务规划；3) 生成逃生路线和辅助导航。因此，应急人员认为 IPS 是减少执勤期间死亡人数、提高规划和现场协调效率的重要工具。
In this survey paper, a systematic, in-depth study of existing schemes for indoor localization of emergency responders was presented. Furthermore, the review was enriched by a tutorial coverage of the localization methods and techniques used for indoor localization. To better understand the specific needs of an IPS for emergency responders, a discussion about the architecture constraints and the specific requirements of such systems was presented. A taxonomy centered on the main design choices to classify the existing IPS for emergency responders was proposed. This taxonomy addresses the technological principle adopted, deployment effort, localization principle, algorithm and environment. A comparison of the different IPS schemes was presented, highlighting the strengths and weaknesses of each approach. Finally, several open issues and future research directions were discussed. 在这篇调查报告中，对现有的应急响应人员室内定位方案进行了系统、深入的研究。此外，还对用于室内定位的定位方法和技术进行了辅导，丰富了综述内容。为了更好地了解应急响应人员对 IPS 的具体需求，还讨论了此类系统的架构限制和具体要求。会上提出了一种分类法，以主要设计选择为中心，对现有的用于应急响应人员的 IPS 进行分类。该分类法涉及所采用的技术原理、部署工作、定位原理、算法和环境。对不同的 IPS 方案进行了比较，强调了每种方法的优缺点。最后，还讨论了几个未决问题和未来的研究方向。
REFERENCES 参考文献
[1] S. He and S.-H. G. Chan, “Wi-Fi fingerprint-based indoor positioning: Recent advances and comparisons,” IEEE Commun. Surveys Tuts., vol. 18, no. 1, pp. 466-490, 1st Quart., 2016. [1] S. He 和 S. -H.G. Chan，"基于 Wi-Fi 指纹的室内定位：Recent advances and comparisons," IEEE Commun.18, no. 1, pp.
[2] (2014). Ubisense. [Online]. Available: www.ubisense.net [2] (2014).Ubisense。[Online].Available: www.ubisense.net
[3] (2014). Sonitor RTL Technologies. [Online]. Available: www.sonitor.com [3] (2014).Sonitor RTL Technologies.[Online].Available: www.sonitor.com
[4] (2014). Ekahau. [Online]. Available: www.ekahau.com [4] (2014).Ekahau.[Online].Available: www.ekahau.com
[5] (2016). Insiteo. Accessed on May 20, 2016. [Online]. Available: http://www.insiteo.com/ [5] (2016).Insiteo。访问日期：2016 年 5 月 20 日。[Online].Available: http://www.insiteo.com/
[6] (2016). Wifarer. Accessed on Jan. 1, 2016. [Online]. Available: http://www.wifarer.com/ [6] (2016).Wifarer.访问日期：2016 年 1 月 1 日。[Online].Available: http://www.wifarer.com/
[7] (2016). Pole Star. Accessed on May 20, 2016. [Online]. Available: https://www.polestar.eu/en/ [7] (2016).极星。访问日期：2016 年 5 月 20 日。[Online].Available: https://www.polestar.eu/en/
[8] (2016). Indoors. Accessed on May 20, 2016. [Online]. Available: http://indoo.rs/ [8] (2016).室内。访问日期：2016 年 5 月 20 日。[Online].Available: http://indoo.rs/
[9] (2016). Infsoft. Accessed on May 20, 2016. [Online]. Available: http://www.infsoft.com/ [9] (2016).Infsoft。访问日期：2016 年 5 月 20 日。[Online].Available: http://www.infsoft.com/
[10] (2016). Navizon. Accessed on May 20, 2016. [Online]. Available: https://www.navizon.com/ [10] (2016).Navizon.访问日期：2016 年 5 月 20 日。[Online].Available: https://www.navizon.com/
[11] (2016). Skyhook. Accessed on May 20, 2016. [Online]. Available: http://www.skyhookwireless.com/ [11] (2016).Skyhook.访问日期：2016 年 5 月 20 日。[Online].Available: http://www.skyhookwireless.com/
[12] C. Fuchs, N. Aschenbruck, P. Martini, and M. Wieneke, “Indoor tracking for mission critical scenarios: A survey,” Pervasive Mobile Comput., vol. 7, no. 1, pp. 1-15, Feb. 2011. [12] C. Fuchs, N. Aschenbruck, P. Martini, and M. Wieneke, "Indoor tracking for mission critical scenarios：普适移动计算》，第 7 卷，第 1 期，第 1-15 页，2011 年 2 月。
[13] L. Faramondi, F. Inderst, F. Pascucci, R. Setola, and U. Delprato, “An enhanced indoor positioning system for first responders,” in Proc. Int. Conf. Indoor Position. Indoor Navig., Montbéliard, France, 2013, pp. 1-8. [13] L. Faramondi、F. Inderst、F. Pascucci、R. Setola 和 U. Delprato，"用于急救人员的增强型室内定位系统"，Proc. Int.Conf.Indoor Position.Indoor Navig.，法国 Montbéliard，2013 年，第 1-8 页。
[14] E. Furey, K. Curran, and P. M. Kevitt, “Probabilistic indoor human movement modeling to aid first responders,” J. Ambient Intell. Humanized Comput., vol. 4, no. 5, pp. 559-569, Oct. 2013. [14] E. Furey, K. Curran, and P. M. Kevitt, "Probabilistic indoor human movement modeling to aid first responders," J. Ambient Intell.Humanized Comput.5, pp.
[15] D. Lymberopoulos et al., “Microsoft indoor localization competition: Experiences and lessons learned,” GetMobile Mobile Comput. Commun., vol. 18, no. 4, pp. 24-31, Oct. 2015. [15] D. Lymberopoulos 等人，"微软室内定位竞赛：经验与教训"，《GetMobile Mobile Comput.Commun., vol. 18, no.4，第 24-31 页，2015 年 10 月。
[16] D. Lymberopoulos et al., “A realistic evaluation and comparison of indoor location technologies,” in Proc. 14th Int. Conf. Inf. Process. Sensor Netw. (IPSN), Seattle, WA, USA, 2015, pp. 178-189. [16] D. Lymberopoulos 等人，"室内定位技术的现实评估与比较"，第 14 届国际会议论文集。Conf.Inf.Process.(IPSN), Seattle, WA, USA, 2015, pp.
[17] D. Harmer et al., “EUROPCOM: Emergency ultrawideband radio for positioning and communications,” in Proc. IEEE Int. Conf. Ultra Wideband, vol. 3. Hanover, Germany, 2008, pp. 85-88. [17] D. Harmer 等人，"EUROPCOM：用于定位和通信的紧急超宽带无线电"，电气和电子工程师学会国际会议论文集。Conf.德国汉诺威，2008 年，第 85-88 页。
[18] M. Femminella and G. Reali, “A zero-configuration tracking system for first responders networks,” IEEE Syst. J., to be published. [18] M. Femminella 和 G. Reali，"用于急救人员网络的零配置跟踪系统"，《IEEE Syst.J.，待出版。
[19] J. Rantakokko, P. Strömbäck, and P. Andersson, “Foot-and kneemounted INS for firefighter localization,” in Proc. Int. Tech. Meeting Inst. Navig., San Diego, CA, USA, 2014, pp. 145-153. [19] J. Rantakokko、P. Strömbäck 和 P. Andersson，"用于消防员定位的脚跪式 INS"，Proc. Int.Tech.会议，美国加利福尼亚州圣迭戈，2014 年，第 145-153 页。
[20] N. Li and B. Becerik-Gerber, “An infrastructure-free indoor localization framework to support building emergency response operations,” in Proc. 19th EG ICE Int. Workshop Intell. Comput. Eng. (ICE), Herrsching, Germany, 2012, pp. 210-219. [20] N. Li and B. Becerik-Gerber, "An infrastructure-free indoor localization framework to support building emergency response operations," in Proc.Workshop Intell.Comput.(ICE), Herrsching, Germany, 2012, pp.
[21] C. L. Fischer, “Localisation, tracking, and navigation support for pedestrians in uninstrumented and unknown environments,” Ph.D. dissertation, School Comput. Commun., Lancaster Univ., Lancashire, U.K., 2012. [21] C. L. Fischer, "Localisation, tracking, and navigation support for pedestrians in uninstrumented and unknown environments," Ph.D. dissertation, School Comput.博士论文，英国兰开斯特大学计算与通信学院，2012 年。
[22] C. Fischer and H. Gellersen, “Location and navigation support for emergency responders: A survey,” IEEE Pervasive Comput., vol. 9, no. 1, pp. 38-47, Jan./Mar. 2010. [22] C. Fischer 和 H. Gellersen，"为应急响应人员提供定位和导航支持：IEEE Pervasive Comput.》，第 9 卷，第 1 期，第 38-47 页，2010 年 1 月/3 月。
[23] M. Algabri, H. Mathkour, H. Ramdane, and M. Alsulaiman, “Comparative study of soft computing techniques for mobile robot navigation in an unknown environment,” Comput. Human Behav., vol. 50, pp. 42-56, Sep. 2015. [23] M. Algabri、H. Mathkour、H. Ramdane 和 M. Alsulaiman，"用于未知环境中移动机器人导航的软计算技术比较研究"，《计算。Human Behav.》，第 50 卷，第 42-56 页，2015 年 9 月。
[24] J. Callmer, “Autonomous localization in unknown environments,” Ph.D. dissertation, Dept. Elect. Eng., Linköping Univ., Linköping, Sweden, 2013. [24] J. Callmer，"未知环境中的自主定位"，博士论文，林雪平大学电子工程系，瑞典林雪平，2013 年。Eng., Linköping Univ., Linköping, Sweden, 2013.
[25] M. Harris, “The way through the flames,” IEEE Spectr., vol. 50, no. 9, pp. 30-35, Sep. 2013. [25] M. Harris，"穿越火焰之路"，《IEEE Spectr》，第 50 卷，第 9 期，第 30-35 页，2013 年 9 月。
[26] G. Glanzer, “Personal and first-responder positioning: State of the art and future trends,” in Proc. Ubiquitous Position. Indoor Navig. Location Based Service (UPINLBS), Helsinki, Finland, 2012, pp. 1-7. [26] G. Glanzer, "Personal and first-responder positioning：State of the art and future trends," in Proc.Ubiquitous Position.Indoor Navig.基于位置的服务（UPINLBS），芬兰赫尔辛基，2012 年，第 1-7 页。
[27] R. Giuliano, F. Mazzenga, M. Petracca, and M. Vari, “Indoor localization system for first responders in emergency scenario,” in Proc. 9th Int. Wireless Commun. Mobile Comput. Conf. (IWCMC), 2013, pp. 1821-1826. [27] R. Giuliano、F. Mazzenga、M. Petracca 和 M. Vari，"紧急情况下第一响应者的室内定位系统"，第 9 届国际无线通信会议。无线通信。Mobile Comput.(IWCMC), 2013, pp.
[28] J. Rantakokko, P. Handel, M. Fredholm, and F. Marsten-Eklof, “User requirements for localization and tracking technology: A survey of mission-specific needs and constraints,” in Proc. Int. Conf. Indoor Position. Indoor Navig., Zürich, Switzerland, 2010, pp. 1-9. [28] J. Rantakokko、P. Handel、M. Fredholm 和 F. Marsten-Eklof，"定位和跟踪技术的用户需求：对特定任务需求和制约因素的调查"，Proc. Int.Conf.Indoor Position.Indoor Navig.，瑞士苏黎世，2010 年，第 1-9 页。
[29] M. Batalin et al., “PHASER: Physiological health assessment system for emergency responders,” in Proc. IEEE Int. Conf. Body Sensor Netw., vol. 6. Cambridge, MA, USA, 2013, pp. 1-6. [29] M. Batalin 等人，"PHASER：应急人员生理健康评估系统"，电气和电子工程师学会国际会议论文集。Conf.美国马萨诸塞州剑桥，2013 年，第 1-6 页。
[30] (2016). Fire Fighter Fatality Investigation and Prevention Program. Accessed on May 25, 2016. [Online]. Available: www.cdc.gov/niosh/fire [30] (2016).消防员死亡调查与预防计划》。访问日期：2016 年 5 月 25 日。[Online].Available: www.cdc.gov/niosh/fire
[31] H. Liu, H. Darabi, P. Banerjee, and J. Liu, “Survey of wireless indoor positioning techniques and systems,” IEEE Trans. Syst., Man, Cybern. C, Appl. Rev., vol. 37, no. 6, pp. 1067-1080, Nov. 2007. [31] H. Liu, H. Darabi, P. Banerjee, and J. Liu, "Survey of wireless indoor positioning techniques and systems," IEEE Trans.Syst.C，Appl. Rev.》，第 37 卷，第 6 期，第 1067-1080 页，2007 年 11 月。
[32] G. Mao, B. Fidan, and B. D. O. Anderson, “Wireless sensor network localization techniques,” Comput. Netw., vol. 51, no. 10, pp. 2529-2553, Jul. 2007. [32] G. Mao, B. Fidan, and B. D. O. Anderson, "Wireless sensor network localization techniques," Comput.51, no. 10, pp.
[33] Y. Gu, A. Lo, and I. Niemegeers, “A survey of indoor positioning systems for wireless personal networks,” Commun. Surveys Tuts., vol. 11, no. 1, pp. 13-32, 1st Quart., 2009. [33] Y. Gu, A. Lo, and I. Niemegeers, "A survey of indoor positioning systems for wireless personal networks," Commun.11, no. 1, pp.
[34] I. Amundson and X. D. Koutsoukos, “A survey on localization for mobile wireless sensor networks,” in Mobile Entity Localization and Tracking in GPS-Less Environnments, R. Fuller and X. D. Koutsoukos, Eds. Heidelberg, Germany: Springer, 2009, pp. 235-254. [34] I. Amundson 和 X. D. Koutsoukos，"A survey on localization for mobile wireless sensor networks"，in Mobile Entity Localization and Tracking in GPS-Less Environnments，R. Fuller and X. D. Koutsoukos，Eds.德国海德堡：Springer, 2009, pp.
[35] D. Zhang, F. Xia, Z. Yang, L. Yao, and W. Zhao, “Localization technologies for indoor human tracking,” in Proc. 5th Int. Conf. Future Inf. Technol., 2010, pp. 1-6. [35] D. Zhang, F. Xia, Z. Yang, L. Yao, and W. Zhao, "Localization technologies for indoor human tracking," in Proc.Conf.Future Inf.2010 年，第 1-6 页。
[36] J. Torres-Solis, T. H. Falk, and T. Chau, “A review of indoor localization technologies: Towards navigational assistance for topographical disorientation,” in Ambient Intelligence, F. J. V. Molina, Ed. Rijeka, Croatia: InTech, 2010, pp. 51-84. [36] J. Torres-Solis、T. H. Falk 和 T. Chau，"室内定位技术综述：Towards navigational assistance for topographical disorientation，" in Ambient Intelligence，F. J. V. Molina，Ed.克罗地亚里耶卡：InTech, 2010, pp.
[37] R. Mautz and S. Tilch, “Survey of optical indoor positioning systems,” in Proc. Int. Conf. Indoor Position. Indoor Navig., Guimarães, Portugal, 2011, pp. 1-7. [37] R. Mautz 和 S. Tilch，"室内光学定位系统调查"，国际会议论文集。Conf.Indoor Navig.Indoor Navig.，葡萄牙吉马良斯，2011 年，第 1-7 页。
[38] A. R. Kulaib, R. M. Shubair, M. A. Al-Qutayri, and J. W. P. Ng, “An overview of localization techniques for wireless sensor networks,” in Proc. Int. Conf. Innov. Inf. Technol., Kuala Lumpur, Malaysia, 2011, pp. 167-172. [38] A. R. Kulaib, R. M. Shubair, M. A. Al-Qutayri, and J. W. P. Ng, "An overview of localization techniques for wireless sensor networks," in Proc. Int.Conf.Innov.Inf.Technol., Kuala Lumpur, Malaysia, 2011, pp.
[39] G. Deak, K. Curran, and J. Condell, “A survey of active and passive indoor localisation systems,” Comput. Commun., vol. 35, no. 16, pp. 1939-1954, Sep. 2012. [39] G. Deak, K. Curran, and J. Condell, "A survey of active and passive indoor localisation systems," Comput.Commun》，第 35 卷，第 16 期，第 1939-1954 页，2012 年 9 月。
[40] S. Yang and Q. Li, “Inertial sensor-based methods in walking speed estimation: A systematic review,” Sensors, vol. 12, no. 6, pp. 6102-6116, May 2012. [40] S. Yang 和 Q. Li，"基于惯性传感器的步行速度估算方法：系统综述》，《传感器》，第 12 卷，第 6 期，第 6102-6116 页，2012 年 5 月。
[41] F. Ijaz, H. K. Yang, A. W. Ahmad, and C. Lee, “Indoor positioning: A review of indoor ultrasonic positioning systems,” in Proc. 15th Int. Conf. Adv. Commun. Technol. (ICACT), 2013, pp. 1146-1150. [41] F. Ijaz、H. K. Yang、A. W. Ahmad 和 C. Lee，"室内定位：15 Int. Conf.Conf.Adv.Commun.Technol.(ICACT), 2013, pp.
[42] R. Harle, “A survey of indoor inertial positioning systems for pedestrians,” IEEE Commun. Surveys Tuts., vol. 15, no. 3, pp. 1281-1293, 3rd Quart., 2013. [42] R. Harle, "A survey of indoor inertial positioning systems for pedestrians," IEEE Commun.Surveys Tuts., vol. 15, no.3, pp.
[43] Z. Farid, R. Nordin, and M. Ismail, “Recent advances in wireless indoor localization techniques and system,” J. Comput. Netw. Commun., vol. 2013, pp. 1-12, Aug. 2013. [43] Z. Farid, R. Nordin, and M. Ismail, "Recent advances in wireless indoor localization techniques and system," J. Comput.Netw.Commun.》，第 2013 卷，第 1-12 页，2013 年 8 月。
[44] G. Han, H. Xu, T. Q. Duong, J. Jiang, and T. Hara, “Localization algorithms of wireless sensor networks: A survey,” Telecommun. Syst., vol. 52, no. 4, pp. 2419-2436, Apr. 2013. [44] G. Han、H. Xu、T. Q. Duong、J. Jiang 和 T. Hara，"无线传感器网络的定位算法：A survey," Telecommun.Syst.4, pp.
[45] H. P. Mistry and N. H. Mistry, “RSSI based localization scheme in wireless sensor networks: A survey,” in Proc. 5th Int. Conf. Adv. Comput. Commun. Technol., Haryana, India, 2015, pp. 647-652. [45] H. P. Mistry 和 N. H. Mistry，"无线传感器网络中基于 RSSI 的定位方案：A survey," in Proc.Conf.Adv.Comput.通信。技术》，印度哈里亚纳邦，2015 年，第 647-652 页。
[46] Z. Yang et al., “Mobility increases localizability: A survey on wireless indoor localization using inertial sensors,” ACM Comput. Surveys, vol. 47, no. 3, pp. 1-34, Apr. 2015. [46] Z. Yang 等人，"移动性提高了定位能力：A survey on wireless indoor locization using inertial sensors," ACM Comput.Surveys, vol. 47, no.3，第 1-34 页，2015 年 4 月。
[47] A. K. M. M. Hossain and W.-S. Soh, “A survey of calibration-free indoor positioning systems,” Comput. Commun., vol. 66, pp. 1-13, Jul. 2015. [47] A. K. M. M. Hossain 和 W.-S.Soh, "A survey of calibration-free indoor positioning systems," Comput.Commun., vol. 66, pp.
[48] D. Dardari, P. Closas, and P. M. Djuric, “Indoor tracking: Theory, methods, and technologies,” IEEE Trans. Veh. Technol., vol. 64, no. 4, pp. 1263-1278, Apr. 2015. [48] D. Dardari、P. Closas 和 P. M. Djuric，"室内跟踪：Theory, methods, and technologies," IEEE Trans.Veh. Technol.Technol.4, pp.
[49] P. Lukowicz, A. Timm-Giel, M. Lawo, and O. Herzog, “WearIT @ work: Toward real-world industrial wearable computing,” IEEE Pervasive Comput., vol. 6, no. 4, pp. 8-13, Oct./Dec. 2007. [49] P. Lukowicz、A. Timm-Giel、M. Lawo 和 O. Herzog，"WearIT @ work：Toward real-world industrial wearable computing," IEEE Pervasive Comput.4, pp.2007.
[50] D. Curone et al., “Smart garments for emergency operators: The ProeTEX project,” IEEE Trans. Inf. Technol. Biomed., vol. 14, no. 3, pp. 694-701, May 2010. [50] D. Curone 等人，"应急操作员的智能服装：The ProeTEX project," IEEE Trans.Inf.Technol.Biomed.3, pp.
[51] N. Li, B. Becerik-Gerber, B. Krishnamachari, and L. Soibelman, “A BIM centered indoor localization algorithm to support building fire emergency response operations,” Autom. Construct., vol. 42, pp. 78-89, Jun. 2014. [51] N. Li, B. Becerik-Gerber, B. Krishnamachari, and L. Soibelman, "A BIM centered indoor localization algorithm to support building fire emergency response operations," Autom.建筑》，第 42 卷，第 78-89 页，2014 年 6 月。
[52] N. Li, Z. Yang, A. Ghahramani, B. Becerik-Gerber, and L. Soibelman, “Situational awareness for supporting building fire emergency response: Information needs, information sources, and implementation requirements,” Fire Safety J., vol. 63, pp. 17-28, Jan. 2014. [52] N. Li、Z. Yang、A. Ghahramani、B. Becerik-Gerber、L. Soibelman，"支持建筑火灾应急响应的态势感知：信息需求、信息来源和实施要求》，《消防安全杂志》，第 63 卷，第 17-28 页，2014 年 1 月。
[53] L. Zhang, S. Alkobaisi, W. D. Bae, and S. Narayanappa, “Ultra wideband indoor positioning system in support of emergency evacuation,” in Proc. 5th ACM SIGSPATIAL Int. Workshop Indoor Spatial Awareness (ISA), Orlando, FL, USA, 2013, pp. 42-49. [53] L. Zhang, S. Alkobaisi, W. D. Bae, and S. Narayanappa, "Ultra wideband indoor positioning system in support of emergency evacuation," in Proc.Workshop Indoor Spatial Awareness (ISA), Orlando, FL, USA, 2013, pp.
[54] P. Pivato, “Analysis and characterization of wireless positioning techniques in indoor environment,” Ph.D. dissertation, Dept. Inf. Eng. Comput. Sci., Univ. at Trento, Trento, Italy, 2012. [54] P. Pivato，"室内环境中无线定位技术的分析和特征描述"，信息工程系博士论文。Eng.Comput.Sci., Univ. at Trento, Trento, Italy, 2012.
[55] K. W. Kolodziej and J. Hjelm, “Sensor systems for indoor position computation,” in Local Positioning Systems: LBS Applications and Services. Boca Raton, FL, USA: Taylor & Francis, 2006, pp. 87-164. [55] K. W. Kolodziej 和 J. Hjelm，"用于室内位置计算的传感器系统"，《本地定位系统》：LBS 应用与服务》。Boca Raton, FL, USA: Taylor & Francis, 2006, pp.
[56] T. K. Sarkar, Z. Ji, K. Kim, A. Medouri, and M. Salazar-Palma, “A survey of various propagation models for mobile communication,” IEEE Antennas Propag. Mag., vol. 45, no. 3, pp. 51-82, Jun. 2003. [56] T. K. Sarkar, Z. Ji, K. Kim, A. Medouri, and M. Salazar-Palma, "A survey of various propagation models for mobile communication," IEEE Antennas Propag.Mag.3, pp.
[57] A. Zanella, “Best practice in RSS measurements and ranging,” IEEE Commun. Surveys Tuts., vol. 18, no. 4, pp. 2662-2686, 4th Quart., 2016. [57] A. Zanella，"RSS 测量和测距的最佳实践"，IEEE Commun.Surveys Tuts., vol. 18, no.4, pp.
[58] A. M. Disha, “A comparative analysis on indoor positioning techniques and systems,” Int. J. Eng. Res. Appl., vol. 3, no. 2, pp. 1790-1796, 2013. [58] A. M. Disha, "A comparative analysis on indoor positioning techniques and systems," Int.J. Eng.Res.Appl.》，第 3 卷，第 2 期，第 1790-1796 页，2013 年。
[59] Z. Yang, Z. Zhou, and Y. Liu, “From RSSI to CSI: Indoor localization via channel response,” ACM Comput. Surveys, vol. 46, no. 2, pp. 1-32, Nov. 2013. [59] Z. Yang, Z. Zhou, and Y. Liu, "From RSSI to CSI: Indoor localization via channel response," ACM Comput.Surveys, vol. 46, no. 2, pp.
[60] X. Wang, L. Gao, S. Mao, and S. Pandey, “DeepFi: Deep learning for indoor fingerprinting using channel state information,” in Proc. IEEE Wireless Commun. Netw. Conf. (WCNC), New Orleans, LA, USA, 2015, pp. 1666-1671. [60] X. Wang、L. Gao、S. Mao 和 S. Pandey，"DeepFi：使用信道状态信息进行室内指纹识别的深度学习"，Proc. IEEE Wireless Commun.Netw.(WCNC), New Orleans, LA, USA, 2015, pp.
[61] K. Wu et al., “CSI-based indoor localization,” IEEE Trans. Parallel Distrib. Syst., vol. 24, no. 7, pp. 1300-1309, Jul. 2013. [61] K. Wu 等人，"基于 CSI 的室内定位"，IEEE Trans.Parallel Distrib.Syst.》，第 24 卷，第 7 期，第 1300-1309 页，2013 年 7 月。
[62] R. E. Kalman, “A new approach to linear filtering and prediction problems,” J. Basic Eng., vol. 82, no. 1, pp. 35-45, 1960. [62] R. E. Kalman，"线性过滤和预测问题的新方法"，《基础工程学报》，第 82 卷，第 1 期，第 35-45 页，1960 年。
[63] J. Hartikainen, A. Solin, and S. Särkkä, “Optimal filtering with Kalman filters and smoothers-A manual for MATLAB toolbox EKF/UKF,” Ph.D. dissertation, Dept. Biomed. Eng. Comput. Sci., Aalto Univ. School Sci., Espoo, Finland, 2011. [63] J. Hartikainen, A. Solin, and S. Särkkä, "Optimal filtering with Kalman filters and smoothers-A manual for MATLAB toolbox EKF/UKF," Ph.D. dissertation, Dept.Eng.Comput.Sci., Aalto Univ. School Sci., Espoo, Finland, 2011.
[64] D. R. Brown and D. B. Dunn, “Classification schemes of positioning technologies for indoor navigation,” in Proc. IEEE Southeastcon, Nashville, TN, USA, 2011, pp. 125-130. [64] D. R. Brown 和 D. B. Dunn，"用于室内导航的定位技术分类方案"，《电气与电子工程师学会东南会议论文集》，美国田纳西州纳什维尔，2011 年，第 125-130 页。
[65] J. Wilson, V. Bhargava, A. Redfern, and P. Wright, “A wireless sensor network and incident command interface for urban firefighting,” in Proc. 4th Annu. Int. Conf. Mobile Ubiquitous Syst. Netw. Services (MobiQuitous), Philadelphia, PA, USA, 2007, pp. 1-7. [65] J. Wilson、V. Bhargava、A. Redfern 和 P. Wright，"用于城市消防的无线传感器网络和事故指挥界面"，Proc. 4th Annu.Int.Conf.Mobile Ubiquitous Syst.Netw.服务（MobiQuitous），美国宾夕法尼亚州费城，2007 年，第 1-7 页。
[66] D. Harmer et al., “An ultra-wide band indoor personnel tracking system for emergency situations (Europcom),” in Proc. Eur. Radar Conf., Amsterdam, The Netherlands, 2008, pp. 404-407. [66] D. Harmer 等人，"用于紧急情况的超宽带室内人员跟踪系统（Europcom）"，Proc.European.阿姆斯特丹，荷兰，2008 年，第 404-407 页。
[67] A. Lo et al., “EUROPCOM-An ultra-wideband (UWB)-based ad hoc network for emergency applications,” in Proc. IEEE Veh. Technol. Conf. VTC Spring, Singapore, 2008, pp. 6-10. [67] A. Lo 等人，"EUROPCOM--基于超宽带（UWB）的紧急应用 ad hoc 网络"，《电气和电子工程师协会车辆技术会议》（Proc. IEEE Veh. Technol.Technol.Conf.VTC Spring, Singapore, 2008, pp.
[68] G. Magenes, D. Curone, M. Lanati, and E. L. Secco, “Long-distance monitoring of physiological and environmental parameters for emergency operators,” in Proc. 31st Annu. Int. Conf. IEEE Eng. Med. Biol. Soc. Eng. Future Biomed. (EMBC), Minneapolis, MN, USA, 2009, pp. 5159-5162. [68] G. Magenes, D. Curone, M. Lanati, and E. L. Secco, "Long-distance monitoring of physiological and environmental parameters for emergency operators," in Proc.Int.Conf.IEEE Eng.Med.Biol.Soc. Eng.Future Biomed.(EMBC), Minneapolis, MN, USA, 2009, pp.
[69] D. Curone et al., “Assessment of sensing fire fighters uniforms for physiological parameter measurement in harsh environment,” IEEE Trans. Inf. Technol. Biomed., vol. 16, no. 3, pp. 501-511, May 2012. [69] D. Curone 等人，"在恶劣环境下感知消防员制服生理参数测量的评估"，IEEE Trans.Inf.Technol.Biomed.3，第 501-511 页，2012 年 5 月。
[70] M. Klann, “Tactical navigation support for firefighters: The lifenet ad-hoc sensor-network and wearable system,” in Mobile Response, vol. 5424. Heidelberg, Germany: Springer, 2009, pp. 41-56. [70] M. Klann，"消防员战术导航支持：lifenet ad-hoc 传感器网络和可穿戴系统"，《移动响应》，第 5424 卷。德国海德堡：Springer, 2009, pp.
[71] S. R. Gandhi, A. Ganz, and G. Mullett, “FIREGUIDE: Firefighter guide and tracker,” in Proc. Annu. Int. Conf. IEEE Eng. Med. Biol., Buenos Aires, Argentina, 2010, pp. 2037-2040. [71] S. R. Gandhi、A. Ganz 和 G. Mullett，"FIREGUIDE：Firefighter guide and tracker," in Proc.Annu.Int.Conf.IEEE Eng.Med.Biol., Buenos Aires, Argentina, 2010, pp.
[72] U. Rüppel, K. M. Stübbe, and U. Zwinger, “Indoor navigation integration platform for firefighting purposes,” in Proc. Int. Conf. Indoor Position. Indoor Navig., Zürich, Switzerland, 2010, pp. 1-6. [72] U. Rüppel, K. M. Stübbe, and U. Zwinger, "Indoor navigation integration platform for firefighting purposes," in Proc. Int.Conf.Indoor Navig.Indoor Navig.，瑞士苏黎世，2010 年，第 1-6 页。
[73] (2016). Identec Solutions. Accessed on Jul. 13, 2016. [Online]. Available: http://www.identecsolutions.com/ [73] (2016).Identec Solutions。访问日期：2016 年 7 月 13 日。[Online].Available: http://www.identecsolutions.com/
[74] G. B. Moon, M. B. Hur, and G.-I. Jee, “An indoor positioning system for a first responder in an emergency environment,” in Proc. 12th Int. Conf. Control Autom. Syst. (ICCAS), 2012, pp. 1368-1372. [74] G. B. Moon、M. B. Hur 和 G.-I.Jee, "An indoor positioning system for a first responder in an emergency environment," in Proc.CONF.Control Autom.Syst.(ICCAS), 2012, pp.
[75] G. B. Moon, S. Chun, M.-B. Hur, and G.-I. Jee, “A robust indoor positioning system using two-stage EKF SLAM for first responders in an emergency environment,” in Proc. 13th Int. Conf. Control Autom. Syst. (ICCAS), 2013, pp. 707-711. [75] G. B. Moon、S. Chun、M.-B.Hur, and G.-I.Jee, "A robust indoor positioning system using two-stage EKF SLAM for first responders in an emergency environment," in Proc.CONF.Control Autom.Syst.(ICCAS), 2013, pp.
[76] S. K. Ghosh, S. Chakraborty, A. Jamthe, and D. P. Agrawal, “Comprehensive monitoring of firefighters by a wireless body area sensor network,” in Proc. 10th Int. Conf. Wireless Opt. Commun. Netw. (WOCN), Gwangju, South Korea, 2013, pp. 1-6. [76] S. K. Ghosh, S. Chakraborty, A. Jamthe, and D. P. Agrawal, "Comprehensive monitoring of firefighters by a wireless body area sensor network," in Proc.Conf.Wireless Opt.Commun.(WOCN), Gwangju, South Korea, 2013, pp.
[77] N. Li, B. Becerik-Gerber, L. Soibelman, and B. Krishnamachari, “Comparative assessment of an indoor localization framework for building emergency response,” Autom. Construct., vol. 57, pp. 42-54, Sep. 2015. [77] N. Li, B. Becerik-Gerber, L. Soibelman, and B. Krishnamachari, "Comparative assessment of an indoor localization framework for building emergency response," Autom.建筑》，第 57 卷，第 42-54 页，2015 年 9 月。
[78] N. Li, B. Becerik-Gerber, and L. Soibelman, “Iterative maximum likelihood estimation algorithm: Leveraging building information and sensing infrastructure for localization during emergencies,” J. Comput. Civil Eng., vol. 26, no. 6, Aug. 2014, Art. no. 4014094. [78] N. Li、B. Becerik-Gerber 和 L. Soibelman，"迭代最大似然估计算法：Leveraging building information and sensing infrastructure for localization during emergencies," J. Comput.土木工程》，第 26 卷，第 6 期，2014 年 8 月，艺术编号：4014094。4014094.
[79] K. Yedavalli and B. Krishnamachari, “Sequence-based localization in wireless sensor networks,” IEEE Trans. Mobile Comput., vol. 7, no. 1, pp. 81-94, Jan. 2008. [79] K. Yedavalli 和 B. Krishnamachari，《无线传感器网络中基于序列的定位》，IEEE Trans.移动计算》，第 7 卷，第 1 期，第 81-94 页，2008 年 1 月。
[80] S. Beauregard, “A helmet-mounted pedestrian dead reckoning system,” in Proc. 3rd Int. Forum Appl. Wearable Comput. (IFAWC), 2006, pp. 1-11. [80] S. Beauregard, "A helmet-mounted pedestrian dead reckoning system," in Proc.可穿戴计算应用论坛（IFAWC），2006 年，第 1-11 页。(IFAWC), 2006, pp.
[81] S. Beauregard, “Omnidirectional pedestrian navigation for first responders,” in Proc. 4th Workshop Position. Navig. Commun., Hanover, Germany, 2007, pp. 33-36. [81] S. Beauregard, "Omnidirectional pedestrian navigation for first responders," in Proc.Navig.通信》，德国汉诺威，2007 年，第 33-36 页。
[82] V. Renaudin, O. Yalak, P. Tomé, and B. Merminod, “Indoor navigation of emergency agents,” Eur. J. Navig., vol. 5, no. 3, pp. 36-45, 2007. [82] V. Renaudin、O. Yalak、P. Tomé 和 B. Merminod，"紧急情况代理的室内导航"，《欧洲导航杂志》，第 5 卷，第 2 期。5, no.3, pp.
[83] L. Ojeda and J. Borenstein, “Non-GPS navigation for security personnel and first responders,” J. Navig., vol. 60, no. 3, pp. 391-407, Sep. 2007. [83] L. Ojeda and J. Borenstein, "Non-GPS navigation for security personnel and first responders," J. Navig.3, pp.
[84] Widyawan, M. Klepal, and S. Beauregard, “A backtracking particle filter for fusing building plans with PDR displacement estimates,” in Proc. 5th Workshop Position. Navig. Commun., 2008, pp. 207-212. [Widyawan, M. Klepal, and S. Beauregard, "A backtracking particle filter for fusing building plans with PDR displacement estimates," in Proc.Navig.通信》，2008 年，第 207-212 页。
[85] O. Woodman and R. Harle, “Pedestrian localisation for indoor environments,” in Proc. 10th Int. Conf. Ubiquitous Comput. (UbiComp), Hanover, Germany, 2008, pp. 114-123. [85] O. Woodman 和 R. Harle，"Pedestrian localisation for indoor environments，" in Proc.Conf.Ubiquitous Comput.(UbiComp), Hanover, Germany, 2008, pp.
[86] B. Cinaz and H. Kenn, “HeadSLAM—Simultaneous localization and mapping with head-mounted inertial and laser range sensors,” in Proc. 12th IEEE Int. Symp. Wearable Comput., Pittsburgh, PA, USA, 2008, pp. 3-10. [86] B. Cinaz 和 H. Kenn，"HeadSLAM--利用头戴式惯性传感器和激光测距传感器进行同步定位和绘图"，第 12 届 IEEE Int. Symp.Wearable Comput.美国宾夕法尼亚州匹兹堡，2008 年，第 3-10 页。
[87] T. Bernoulli, G. Glanzer, T. Wießflecker, R. Schütz, and U. Walder, “A building information model for a context-adaptive disaster management system,” in Proc. Intell. Comput. Eng. (ICE), 2008, pp. 402-409. [87] T. Bernoulli, G. Glanzer, T. Wießflecker, R. Schütz, and U. Walder, "A building information model for a context-adaptive disaster management system," in Proc. Intell.Comput.(ICE), 2008, pp.
[88] U. Walder and T. Wießflecker, “An indoor positioning system for improved action force command and disaster management,” in Proc. 6th Int. ISCRAM Conf., 2009, pp. 251-262. [88] U. Walder 和 T. Wießflecker, "An indoor positioning system for improved action force command and disaster management," in Proc.ISCRAM Conf.，2009 年，第 251-262 页。
[89] T. Bernoulli, G. Glanzer, T. Weisflecker, and U. Walder, “Infrastructurless indoor positioning system for first responders,” in Proc. 7th Int. ISCRAM Conf., 2010, pp. 1-10. [89] T. Bernoulli, G. Glanzer, T. Weisflecker, and U. Walder, "Infrastructurless indoor positioning system for first responders," in Proc.ISCRAM Conf., 2010, pp.
[90] J. Rantakokko et al., “Accurate and reliable soldier and first responder indoor positioning: Multisensor systems and cooperative localization,” IEEE Wireless Commun., vol. 18, no. 2, pp. 10-18, Apr. 2011. [90] J. Rantakokko 等人，"准确可靠的士兵和急救人员室内定位：18, no. 2, pp.
[91] J. Callmer, D. Törnqvist, and F. Gustafsson, “Probabilistic stand still detection using foot mounted IMU,” in Proc. 13th Int. Conf. Inf. Fusion, Edinburgh, U.K., 2010, pp. 1-7. [91] J. Callmer、D. Törnqvist 和 F. Gustafsson，"使用安装在脚上的 IMU 进行概率静止检测"，第 13 届国际计算机大会论文集。Conf.Inf.Fusion, Edinburgh, U.K., 2010, pp.
[92] R. Zhang, F. Hoflinger, and L. Reindl, “Inertial sensor based indoor localization and monitoring system for emergency responders,” IEEE Sensors J., vol. 13, no. 2, pp. 838-848, Feb. 2013. [92] R. Zhang, F. Hoflinger, and L. Reindl, "Inertial sensor based indoor locization and monitoring system for emergency responders," IEEE Sensors J., vol. 13, no. 2, pp.
[93] R. Zhang, M. Loschonsky, and L. M. Reindl, “Study of zero velocity update for both low-and high-speed human activities,” Int. J. E-Health Med. Commun., vol. 2, no. 2, pp. 46-67, Apr. 2011. [93] R. Zhang, M. Loschonsky, and L. M. Reindl, "Study of zero velocity update for both low and high-speed human activities," Int. J. E-Health Med.J. E-Health Med.通讯》，第 2 卷第 2 期，第 46-67 页，2011 年 4 月。
[94] K. V. S. Hari et al., “A prototype of a first-responder indoor localization system,” J. Indian Inst. Sci., vol. 93, no. 3, pp. 511-520, 2013. [94] K. V. S. Hari 等人，"A prototype of a first-responder indoor localization system"，J. Indian Inst. Sci.，vol. 93，no.3, pp.
[95] (2016). OpenShoe. Accessed on May 19, 2016. [Online]. Available: http://www.openshoe.org/ [95] (2016).OpenShoe.访问日期：2016 年 5 月 19 日。[Online].Available: http://www.openshoe.org/
[96] J.-O. Nilsson, I. Skog, P. Händel, and K. V. S. Hari, “Foot-mounted INS for everybody-An open-source embedded implementation,” in Proc. IEEE/ION Position Location Navig. Symp., Myrtle Beach, SC, USA, 2012, pp. 140-145. [96] J.-O.Nilsson, I. Skog, P. Händel, and K. V. S. Hari, "Foot-mounted INS for everybody-An open-source embedded implementation," in Proc.Symp., Myrtle Beach, SC, USA, 2012, pp.
[97] I. Skog, J.-O. Nilsson, D. Zachariah, and P. Händel, “Fusing the information from two navigation systems using an upper bound on their maximum spatial separation,” in Proc. Int. Conf. Indoor Position. Indoor Navig. (IPIN), Sydney, NSW, Australia, 2012, pp. 1-5. [97] I. Skog, J.-O.Nilsson, D. Zachariah, and P. Händel, "Fusing the information from two navigation systems using an upper bound on their maximum spatial separation," in Proc.Conf.Indoor Position.(IPIN), Sydney, NSW, Australia, 2012, pp.
[98] J.-O. Nilsson, D. Zachariah, I. Skog, and P. Händel, “Cooperative localization by dual foot-mounted inertial sensors and inter-agent ranging,” EURASIP J. Adv. Signal Process., vol. 2013, no. 1, pp. 1-17, 2013. [98] J.-O.Nilsson, D. Zachariah, I. Skog, and P. Händel, "Cooperative localization by dual foot-mounted inertial sensors and inter-agent ranging," EURASIP J. Adv. Signal Process.
[99] F. Pascucci et al., “A REference implementation of interoperable indoor location & communication systems for first REsponders: The REFIRE project,” in Proc. IEEE Int. Symp. Safety Security Rescue Robot. (SSRR), College Station, TX, USA, 2012, pp. 1-5. [99] F. Pascucci 等人，"A REference implementation of interoperable indoor location & communication systems for first REsponders：The REFIRE project," in Proc.Symp.Safety Security Rescue Robot.(SSRR), College Station, TX, USA, 2012, pp.
[100] J.-O. Nilsson et al., “Accurate indoor positioning of firefighters using dual foot-mounted inertial sensors and inter-agent ranging,” in Proc. IEEE/ION Position Location Navig. Symp. (PLANS), Monterey, CA, USA, 2014, pp. 631-636. [100] J.-O.Nilsson 等人，"利用双脚安装的惯性传感器和机构间测距对消防员进行精确室内定位"，《电气和电子工程师学会/国际集成电路定位导航研讨会》（Proc. IEEE/ION Position Location Navig.Symp.(PLANS), Monterey, CA, USA, 2014, pp.
[101] C. Fischer, K. Muthukrishnan, M. Hazas, and H. Gellersen, “Ultrasound-aided pedestrian dead reckoning for indoor navigation,” in Proc 1st ACM Int. Workshop Mobile Entity Localization Tracking GPS Less Environ. (MELT), San Francisco, CA, USA, 2008, pp. 31-36. [101] C. Fischer, K. Muthukrishnan, M. Hazas, and H. Gellersen, "Ultrasound-aided pedestrian dead reckoning for indoor navigation," in Proc 1st ACM Int. Workshop Mobile Entity Localization Tracking GPS Less Environ.(MELT), San Francisco, CA, USA, 2008, pp.
[102] C. Fischer, P. T. Sukumar, and M. Hazas, “Tutorial: Implementing a pedestrian tracker using inertial sensors,” IEEE Pervasive Comput., vol. 12, no. 2, pp. 17-27, Apr./Jun. 2013. [102] C. Fischer、P. T. Sukumar 和 M. Hazas，"教程：使用惯性传感器实现行人跟踪器"，《IEEE Pervasive Comput》，第 12 卷，第 2 期，第 17-27 页，2013 年 4 月/6 月。
[103] V. Amendolare et al., “WPI precision personnel locator system: Inertial navigation supplementation,” in Proc. IEEE/ION Position Location Navig. Symp., 2008, pp. 350-357. [103] V. Amendolare 等人，"WPI 精确人员定位系统：Inertial navigation supplementation," in Proc.Symp., 2008, pp.
[104] A. Cavanaugh, M. Lowe, D. Cyganski, and R. J. Duckworth, “A Bayesian fusion algorithm for precision personnel location in indoor environments,” in Proc. Int. Tech. Meeting Inst. Navig., San Diego, CA, USA, 2011, pp. 397-403. [104] A. Cavanaugh, M. Lowe, D. Cyganski, and R. J. Duckworth, "A Bayesian fusion algorithm for precision personnel location in indoor environments," in Proc. Int.技术会议会议，美国加利福尼亚州圣迭戈，2011 年，第 397-403 页。
[105] A. Cavanaugh, M. Lowe, D. Cyganski, and R. J. Duckworth, “WPI precision personnel location system: Rapid deployment antenna system and sensor fusion for 3D precision location,” in Proc. Int. Tech. Meeting Inst. Navig., San Diego, CA, USA, 2010, pp. 214-219. [105] A. Cavanaugh、M. Lowe、D. Cyganski 和 R. J. Duckworth，"WPI 精确人员定位系统：Rapid deployment antenna system and sensor fusion for 3D precision location," in Proc.技术会议会议，美国加利福尼亚州圣迭戈，2010 年，第 214-219 页。
[106] B. Woodacre, D. Cyganski, R. J. Duckworth, and V. Amendolare, “WPI precision personnel locator system: Antenna geometry estimation using a robust multilateralization technique,” in Proc. Int. Tech. Meeting Inst. Navig., Anaheim, CA, USA, 2009, pp. 822-828. [106] B. Woodacre、D. Cyganski、R. J. Duckworth 和 V. Amendolare，"WPI 精确人员定位系统：使用稳健的多边化技术进行天线几何估算"，Proc. Int.Tech.会议，美国加利福尼亚州阿纳海姆，2009 年，第 822-828 页。
[107] J. Coyne, D. Cyganski, and R. J. Duckworth, “FPGA-based coprocessor for singular value array reconciliation tomography,” in Proc. 16th Int. Symp. Field Program. Custom Comput. Mach., Palo Alto, CA, USA, 2008, pp. 163-172. [107] J. Coyne、D. Cyganski 和 R. J. Duckworth，"基于 FPGA 的奇异值阵列调和断层摄影协处理器"，第 16 届国际会议。Symp.Field Program.Custom Comput.2008 年，美国加利福尼亚州帕洛阿尔托，第 163-172 页。
[108] Widyawan et al., “Virtual lifeline: Multimodal sensor data fusion for robust navigation in unknown environments,” Pervasive Mobile Comput., vol. 8, no. 3, pp. 388-401, Jun. 2012. [108] Widyawan 等人，"虚拟生命线：多模态传感器数据融合实现未知环境中的稳健导航"，《普适移动计算》，第 8 卷第 3 期，第 388-401 页，2012 年 6 月。3, pp.
[109] W. Hawkinson et al., “GLANSER: Geospatial location, accountability, and navigation system for emergency responders-system concept and performance assessment,” in Proc. IEEE/ION Position Location Navig. Symp., 2012, pp. 98-105. [109] W. Hawkinson 等人，"GLANSER：Geospatial location, accountability, and navigation system for emergency responders-system concept and performance assessment，"in Proc. IEEE/ION Position Location Navig.Symp., 2012, pp.
[110] S. Huseth, B. Dewberry, and R. McCroskey, “Pulsed-RF ultrawideband ranging for the GLANSER GPS-denied emergency responder navigation system,” in Proc. Int. Tech. Meeting Inst. Navig., San Diego, CA, USA, 2011, pp. 389-396. [110] S. Huseth、B. Dewberry 和 R. McCroskey，"GLANSER 全球定位系统拒绝应急响应导航系统的脉冲射频超宽带测距"，Proc. Int.技术会议会议，美国加利福尼亚州圣迭戈，2011 年，第 389-396 页。
[111] L. Haas and M. Harlacher, “First responder location and tracking using synthetic aperture navigation,” in Proc. IEEE/ION Position Location Navig. Symp. (PLANS), Indian Wells, CA, USA, 2010, pp. 483-487. [111] L. Haas 和 M. Harlacher，"First Responder location and tracking using synthetic aperture navigation"，in Proc. IEEE/ION Position Location Navig.Symp.(PLANS), Indian Wells, CA, USA, 2010, pp.
[112] (2016). WASP—Wearable Advanced Sensor Platform. Accessed on Jan. 27, 2016. [Online]. Available: http://globeturnoutgear.com/ innovation/wasp [112] (2016).WASP--可穿戴先进传感器平台。2016年1月27日访问。[Online].网址：http://globeturnoutgear.com/ innovation/wasp
[113] (2016). TRX NEON Systems. Accessed on Jan. 27, 2016. [Online]. Available: http://www.trxsystems.com/ [113] (2016).TRX NEON 系统。访问日期：2016 年 1 月 27 日。[Online].Available: http://www.trxsystems.com/
[114] N. Simon et al., “Indoor localization system for emergency responders with ultra low-power radio landmarks,” in Proc. IEEE Int. Instrum. Meas. Technol. Conf. (I2MTC), Pisa, Italy, 2015, pp. 309-314. [114] N. Simon 等人，"使用超低功耗无线电地标的应急响应人员室内定位系统"，《电气和电子工程师学会国际会议论文集》（Proc. IEEE Int. Instrum.Instrum.Meas.Technol.(I2MTC), Pisa, Italy, 2015, pp.
[115] F. Hoflinger, R. Zhang, and L. M. Reindl, “Indoor-localization system using a micro-inertial measurement unit (IMU),” in Proc. Eur. Freq. Time Forum, 2012, pp. 443-447. [115] F. Hoflinger、R. Zhang 和 L. M. Reindl，"使用微型惯性测量单元（IMU）的室内定位系统"，Proc.Eur.Freq.时间论坛，2012 年，第 443-447 页。
[116] (2014). Summit Safety—The Technology Company. [Online]. Available: http://summitsafetyinc.com/ [116] (2014).高峰安全技术公司。[Online].Available: http://summitsafetyinc.com/
[117] J. Blankenbach, A. Norrdine, and H. Hellmers, “A robust and precise 3D indoor positioning system for harsh environments,” in Proc. Int. Conf. Indoor Position. Indoor Navig. (IPIN), Sydney, NSW, Australia, 2012, pp. 1-8. [117] J. Blankenbach, A. Norrdine, and H. Hellmers, "A robust and precise 3D indoor positioning system for harsh environments," in Proc.Conf.Indoor Position.(IPIN), Sydney, NSW, Australia, 2012, pp.
[118] J. Blankenbach, A. Norrdine, H. Hellmers, and E. Gasparian, “A novel magnetic indoor positioning system for indoor location services,” in Proc. 8th Int. Symp. Location Based Services, Vienna, Austria, 2011, pp. 1-11. [118] J. Blankenbach, A. Norrdine, H. Hellmers, and E. Gasparian, "A novel magnetic indoor positioning system for indoor location services," in Proc.Symp.基于位置的服务》，奥地利维也纳，2011 年，第 1-11 页。
[119] J. Blankenbach, A. Norrdine, and H. Hellmers, “Adaptive signal processing for a magnetic indoor positioning system,” in Proc. Indoor Position. Indoor Navig., 2011, pp. 1-4. [119] J. Blankenbach, A. Norrdine, and H. Hellmers, "Adaptive signal processing for a magnetic indoor positioning system," in Proc.室内导航》，2011 年，第 1-4 页。
[120] J. Blankenbach and A. Norrdine, “Position estimation using artificial generated magnetic fields,” in Proc. Int. Conf. Indoor Position. Indoor Navig., 2010, pp. 1-5. [120] J. Blankenbach 和 A. Norrdine，"使用人工生成的磁场进行位置估算"，国际会议论文集。Conf.Indoor Position.Indoor Navig.，2010 年，第 1-5 页。
[121] (2016). ProFiTex—Advanced Protective Firefighting Equipment. Accessed on Jan. 27, 2016. [Online]. Available: https://www.ims.tuwien.ac.at/projects/profitex [121] (2016).ProFiTex-高级防护消防设备。访问日期：2016 年 1 月 27 日。[Online].Available: https://www.ims.tuwien.ac.at/projects/profitex
[122] C. Schönauer, E. Vonach, G. Gerstweiler, and H. Kaufmann, “3D building reconstruction and thermal mapping in fire brigade operations,” in Proc. IEEE Virtual Reality (VR), Lake Buena Vista, FL, USA, 2013, pp. 1-2. [122] C. Schönauer、E. Vonach、G. Gerstweiler 和 H. Kaufmann，"消防队行动中的 3D 建筑重建和热映射"，《电气和电子工程师学会虚拟现实（VR）会议》，美国佛罗里达州布纳维斯塔湖，2013 年，第 1-2 页。
[123] T. Van Haute et al., “Comparability of RF-based indoor localization solutions in heterogeneous environments: An experimental study,” Int. J. Ad Hoc Ubiquitous Comput., vol. 1, no. 1, pp. 1-23, 2015. [123] T. Van Haute 等人，"异构环境中基于射频的室内定位解决方案的可比性：An experimental study," Int.Ad Hoc Ubiquitous Comput.》，第 1 卷第 1 期，第 1-23 页，2015 年。
[124] T. Van Haute et al., “Platform for benchmarking of RF-based indoor localization solutions,” IEEE Commun. Mag., vol. 53, no. 9, pp. 126-133, Sep. 2015. [124] T. Van Haute 等人，"基于射频的室内定位解决方案基准测试平台"，IEEE Commun.杂志》，第 53 卷，第 9 期，第 126-133 页，2015 年 9 月。
[125] (2016). NIST Smart Firefighting Project. Accessed on Feb. 1, 2016. [Online]. Available: http://www.nist.gov/el/fire_research/firetech/ project_sff.cfm [125] (2016).NIST 智能消防项目。访问日期：2016 年 2 月 1 日。[Online].网址：http://www.nist.gov/el/fire_research/firetech/ project_sff.cfm
[126] Z. Wang et al., “Cyber-physical systems for water sustainability: Challenges and opportunities,” IEEE Commun. Mag., vol. 53, no. 5, pp. 216-222, May 2015. [126] Z. Wang 等人，"水可持续性的网络物理系统：Challenges and opportunities," IEEE Commun.Mag.5，第 216-222 页，2015 年 5 月。
[127] X. Zhang, Y. Jin, H.-X. Tan, and W.-S. Soh, “CIMLoc: A crowdsourcing indoor digital map construction system for localization,” in Proc. IEEE 9th Int. Conf. Intell. Sensors Sensor Netw. Inf. Process. (ISSNIP), Singapore, 2014, pp. 1-6. [127] X. Zhang, Y. Jin, H.-X.Tan, and W.-S.Soh, "CIMLoc: A crowdsourcing indoor digital map construction system for localization," in Proc.Conf.Intell.Sensors Sensor Netw.Inf.Process.(ISSNIP), Singapore, 2014, pp.
[128] Y. Gu and F. Ren, “Energy-efficient indoor localization of smart handheld devices using bluetooth,” IEEE Access, vol. 3, pp. 1450-1461, 2015. [128] Y. Gu 和 F. Ren，"利用蓝牙实现智能手持设备的高能效室内定位"，《IEEE Access》，第 3 卷，第 1450-1461 页，2015 年。
[129] H. Xie et al., “Accelerating crowdsourcing based indoor localization using CSI,” in Proc. IEEE 21st Int. Conf. Parallel Distrib. Syst. (ICPADS), Melbourne, VIC, Australia, 2015, pp. 274-281. [129] H. Xie 等人，"利用 CSI 加速基于众包的室内定位"，《第 21 届 IEEE 国际会议》（Proc. IEEE 21st Int.Conf.Parallel Distrib.Syst.(ICPADS), Melbourne, VIC, Australia, 2015, pp.
André Filipe Gonçalves Ferreira was born in Braga, Portugal, in 1988. He received the integrated master’s degree in electronic engineering from the University of Minho, Guimarães, Portugal, in 2012, where he is currently pursuing the Ph.D. degree with the Department of Electronics at Algoritmi Center. His research interests are in 3-D indoor localization and communication in unknown and unstructured environments, wireless body sensor networks for emergency responders, and field-programmable gate array-based embedded systems. He was a recipient of a Ph.D. scholarship from the Portuguese Foundation of Science and Technology. He is conducting research on the field of 3-D indoor localization for emergency responders in unknown and unstructured environments. André Filipe Gonçalves Ferreira 于 1988 年出生于葡萄牙布拉加。他于 2012 年获得葡萄牙吉马良斯米尼奥大学电子工程综合硕士学位，目前正在 Algoritmi 中心电子系攻读博士学位。他的研究兴趣是未知和非结构化环境中的三维室内定位和通信、用于应急响应人员的无线人体传感器网络以及基于现场可编程门阵列的嵌入式系统。他曾获得葡萄牙科技基金会的博士奖学金。他正在未知和非结构化环境中为应急响应人员进行三维室内定位领域开展研究。
Duarte Manuel Azevedo Fernandes was born in V. N. Famalicão, Portugal, in 1987. He received the integrated master’s degree in electronic engineering from the University of Minho, Guimarães, Portugal, in 2012, where he is currently pursuing the Ph.D. degree with the Department of Electronics at Algoritmi Center. He was a recipient of a Ph.D. scholarship from the Portuguese Foundation of Science and Technology. He is conducting research on the field of wireless body sensor networks for emergency responders. His research interests are in wireless body sensor network, reliable and energy efficient communication, transmission power control mechanisms, opportunistic packet scheduling mechanisms, and 3-D indoor localization and communication in unknown and unstructured environments. 杜阿尔特-曼努埃尔-阿泽维多-费尔南德斯 1987 年出生于葡萄牙 V. N. Famalicão。他于 2012 年获得葡萄牙吉马良斯米尼奥大学电子工程综合硕士学位，目前正在 Algoritmi 中心电子系攻读博士学位。他曾获得葡萄牙科技基金会的博士奖学金。他正在为应急响应人员进行无线人体传感器网络领域的研究。他的研究兴趣是无线人体传感器网络、可靠和节能通信、传输功率控制机制、机会主义数据包调度机制以及未知和非结构化环境下的三维室内定位和通信。
André Paulo Catarino received the degree in electrical and computers engineering from the Faculty of Engineering, University of Porto in 1992, and the M.Sc. and Ph.D. degrees in the textile engineering and technology areas from the School of Engineering in 1998 and 2005, respectively. Since 1994, he has been with the Textile Engineering Department, University of Minho, where he is currently an Assistant Professor. He is a Permanent Researcher with the Center of Textile Science and Technology. His specific interests lie on knitting technology and processes and interactive textiles, especially, applied on health and sports. André Paulo Catarino 于 1992 年获得波尔图大学工程学院电子和计算机工程学位，并于 1998 年和 2005 年分别获得工程学院纺织工程和技术领域的理学硕士和博士学位。自 1994 年以来，他一直在米尼奥大学纺织工程系工作，现任助理教授。他还是纺织科学与技术中心的长期研究员。他的主要研究方向是针织技术和工艺以及交互式纺织品，尤其是应用于健康和运动领域的交互式纺织品。
João L. Monteiro (M’92) received the B.S. degree in electrical engineering from the University of Porto, Portugal, in 1980, and the Ph.D. degree in computer engineering from the University of Minho, Portugal, in 1991. He is currently a Full Professor with the Industrial Electronics Department, University of Minho. His main research interests include hardware/software codesign, automatic development of discrete-event controllers, sensors and distributed systems for building supervision, management, and control (“domotic” or intelligent buildings), industrial networking and distributed systems for online supervision and management, and signal processing. He supervises several Ph.D. students and coordinates teams at the University of Minho for several research and development projects. Since 2013, he has been the Dean of the School of Engineering. João L. Monteiro（92 年硕士）于 1980 年获得葡萄牙波尔图大学电气工程学士学位，1991 年获得葡萄牙米尼奥大学计算机工程博士学位。他目前是米尼奥大学工业电子系的全职教授。他的主要研究方向包括硬件/软件代码设计、离散事件控制器的自动开发、用于楼宇监督、管理和控制的传感器和分布式系统（"domotic "或智能楼宇）、用于在线监督和管理的工业网络和分布式系统以及信号处理。他指导多名博士生，并协调米尼奥大学多个研发项目团队的工作。自 2013 年以来，他一直担任工程学院院长。
Manuscript received October 17, 2016; revised March 31, 2017; accepted May 8, 2017. Date of publication May 12, 2017; date of current version November 21, 2017. This work was supported in part by the Foundation for Science and Technology (FCT)—within the Project Scope under Grant Pest-OE/EEI/UI0319/2014, in part by the FEDER funds through the Competitively Factor Operational Programmed-COMPETE, and in part by the national founds through FCT within the scope under Project POCI-01-0145-FEDER-007136. The work of A. F. G. Ferreira and D. M. A. Fernandes was supported by the FCT under Grant SFRH/BD/91477/2012 and Grant SFRH/BD/92082/2012. (Corresponding author: André Filipe Gonçalves Ferreira.) 2016年10月17日收到稿件；2017年3月31日修改；2017年5月8日接受。发表日期：2017年5月12日；当前版本日期：2017年11月21日。这项工作部分得到了科学技术基金会（FCT）的支持--在 Pest-OE/EEI/UI0319/2014 号补助金的项目范围内；部分得到了 FEDER 基金通过竞争性因素运作计划--COMPETE 的支持；部分得到了国家基金会通过 FCT 在 POCI-01-0145-FEDER-007136 号项目范围内的支持。费雷拉（A. F. G. Ferreira）和费尔南德斯（D. M. A. Fernandes）的工作得到了 FCT SFRH/BD/91477/2012 号赠款和 SFRH/BD/92082/2012 号赠款的支持。(通讯作者：安德烈-菲利佩-贡萨尔维斯-费雷拉）。
A. F. G. Ferreira, D. M. A. Fernandes, and J. L. Monteiro are with the Department of Industrial Electronics, University of Minho, 4800-058 Guimarães, Portugal (e-mail: id4541@alunos.uminho.pt; id4542@alunos.uminho.pt; joao.monteiro@dei.uminho.pt). A. F. G. Ferreira、D. M. A. Fernandes 和 J. L. Monteiro 现任葡萄牙米尼奥大学工业电子系，邮编 4800-058 Guimarães（电子邮箱：id4541@alunos.uminho.pt；id4542@alunos.uminho.pt；joao.monteiro@dei.uminho.pt）。
A. P. Catarino is with the Center of Textile Science and Technology, University of Minho, 4800-058 Guimarães, Portugal (e-mail: whiteman@det.uminho.pt). A. P. Catarino现供职于葡萄牙米尼奥大学纺织科学与技术中心，地址：4800-058 Guimarães（电子邮箱：whiteman@det.uminho.pt）。
Digital Object Identifier 10.1109/COMST.2017.2703620 数字对象标识符 10.1109/COMST.2017.2703620
^(1){ }^{1} The terms landmark and waypoint are used to describe devices or structures that are used to correct the target’s position as it enters a building. They differ from each other on the position information they provide. A landmark provides an absolute position with high accuracy and its deployment occurs beforehand. The waypoint position is computed in real time, at the same time that the target position is being computed. A waypoint provides the relative position with an error of the same magnitude of the IPS associated error, its deployment is on-the-fly. ^(1){ }^{1} "地标 "和 "航点 "这两个术语用于描述在目标进入建筑物时用于修正目标位置的装置或结构。它们提供的位置信息各不相同。地标提供的是高精度的绝对位置，其部署是事先进行的。航点位置是在计算目标位置的同时实时计算的。航点提供相对位置，其误差与 IPS 相关误差的大小相同，其部署是即时进行的。
