这是用户在 2025-3-10 14:30 为 https://app.immersivetranslate.com/pdf-pro/56570de3-f5d2-47e4-aa23-550546b3b3b8/?isTrial=true 保存的双语快照页面，由 沉浸式翻译 提供双语支持。了解如何保存？

Privacy-Preserving AGV Collision-Resistance at the Edge Using Location-Based Encryption
基于位置加密的隐私保护AGV边缘防撞

Liang Zhang ⊕ ⊕ ^(o+){ }^{\oplus}, Haibin Kan ( ( ^((){ }^{( }, Member, IEEE, and Yihao Wang ( ( ^((){ }^{( }
张亮 ⊕ ⊕ ^(o+){ }^{\oplus} ，靳海滨 ( ( ^((){ }^{( } ，IEEE会员，王一浩 ( ( ^((){ }^{( }

Abstract  摘要

Edge computing fundamentally changes the architecture in which applications are deployed and how resources are managed, especially in the 5G/6G era. Automated guided vehicles (AGV) are critical means of transportation in future society. We aim at investigating a privacy-preserving AGV collision-resistance paradigm leveraging edge computing. We define a model containing all potential collision occasions and propose a method to handle collisions using “virtual” traffic lights generated by edge servers. The proposed paradigm is AGV-centered, for AGVs ought not to care about where the nearby edge servers are in a broadcasting environment. In case of privacy leakage, we incorporate ciphertext-policy attribute-based encryption (CP-ABE) to protect AGV information. Besides, the use of CP-ABE leads to a location-based encryption (LBE) scheme, where AGV encrypts its running state information leveraging its current position rather than an edge server’s public key. That is achievable due to the support of integer comparison in the CP-ABE access policy. To our knowledge, we are the first to implement an LBE scheme based on CP-ABE. Part of the paradigm has been deployed in a wharf yard in Shanghai. Throughout the article, we take a yard as an example and conduct concrete experiments to highlight the feasibility and efficiency.
边缘计算从根本上改变了部署应用程序的架构以及管理资源的方式，特别是在5G/6 G时代。自动导引车（AGV）是未来社会的重要交通工具。我们的目标是研究利用边缘计算的隐私保护AGV抗碰撞范例。我们定义了一个包含所有潜在碰撞情况的模型，并提出了一种使用边缘服务器生成的“虚拟”交通灯来处理碰撞的方法。所提出的范例是以AGV为中心的，因为AGV不应该关心附近的边缘服务器在广播环境中的位置。在隐私泄露的情况下，我们将密文策略基于属性的加密（CP-ABE）来保护AGV信息。此外，使用CP-ABE导致基于位置的加密（LBE）方案，其中AGV利用其当前位置而不是边缘服务器的公钥来加密其运行状态信息。 由于CP-ABE访问策略中对整数比较的支持，这是可以实现的。据我们所知，我们是第一个实现基于CP-ABE的LBE方案。部分范例已部署在上海的码头堆场。在本文中，我们以一个堆场为例，进行了具体的实验，以突出的可行性和效率。

Index Terms-AGV, collision-resistance, CP-ABE, location-based encryption, edge computing, privacy-preserving
索引术语-AGV，抗碰撞，CP-ABE，基于位置的加密，边缘计算，隐私保护

1 Introduction  1引言

IN recent years, the development of industrial internet-ofthings (IIoT) has gained significant momentum, heralding a variety of associated techniques and solutions, such as data management [1], security and privacy [2], [3], blockchain [4], [5] and data mining [6]. 5G is an increasingly astute innovation that interconnects the whole society by the massive IoT objects [7]. As a core technique, edge computing [8] allows applications to be deployed on edge
近年来，工业物联网（IIoT）的发展势头强劲，预示着各种相关技术和解决方案，如数据管理[1]，安全和隐私[2]，[3]，区块链[4]，[5]和数据挖掘[6]。5G是一项日益精明的创新，它通过大量的物联网对象将整个社会互联起来[7]。作为核心技术，边缘计算[8]允许将应用程序部署在边缘上
  • Liang Zhang is with the School of Cyberspace Security(School of Cryptology), Hainan University, Haikou, Hainan 570228, China, and also with the School of Computer Science, Fudan University, Shanghai 200433, China. E-mail: zhangliang@hainanu.edu.cn.
    张亮就职于海南大学网络安全学院（密码学学院），海南海口570228，中国，复旦大学计算机科学学院，上海200433，中国。电子邮件地址：zhangliang@hainanu.edu.cn。
  • Haibin Kan is with the Shanghai Key Laboratory of Intelligent Information Processing, School of Computer Science, Fudan University, Shanghai 200433, China, and with the Shanghai Engineering Research Center of Blockchain, Shanghai 200433, China, and also with Zhuhai Fudan Innovation Research Institute, Zhuhai, Guangdong 518057, China. E-mail: hbkan@fudan.edu.cn.
    Kan Haibin是复旦大学计算机学院上海市智能信息处理重点实验室，上海200433，中国，上海区块链工程研究中心，上海200433，中国，珠海复旦创新研究院，广东珠海518057，中国。电子邮件：hbkan@fudan.edu.cn。
  • Yihao Wang is with the Shanghai Key Laboratory of Intelligent Information Processing, School of Computer Science, Fudan University, Shanghai 200433, China, and also with the Shanghai Engineering Research Center of Blockchain, Shanghai 200433, China. E-mail: 20210240262 @ f u d a n . e d u . c n 20210240262 @ f u d a n . e d u . c n 20210240262@fudan.edu.cn20210240262 @ f u d a n . e d u . c n.
    Yihao Wang是复旦大学计算机科学学院上海市智能信息处理重点实验室，上海200433，中国，也是上海区块链工程研究中心，上海200433，中国。电子邮件： 20210240262 @ f u d a n . e d u . c n 20210240262 @ f u d a n . e d u . c n 20210240262@fudan.edu.cn20210240262 @ f u d a n . e d u . c n 。
Manuscript received 18 August 2022; revised 26 December 2022; accepted 4 January 2023. Date of publication 9 January 2023; date of current version 8 August 2023.
2022年8月18日收到Mandarin pt; 2022年12月26日修订; 2023年1月4日接受。发布日期2023年1月9日;当前版本日期2023年8月8日。
This work was supported in part by the National Key R&D Program of China under Grant 2019YFB2101703, in part by the Foundation of Hainan University under Grant KYQD22168, in part by the Key Laboratory of Internet Information Retrieval of Hainan Province Research Found under Grant 2022KY01, in part by the National Natural Science Foundation of China under Grants 62272107 and U19A2066, in part by the Innovation Action Plan of Shanghai Science and Technology under Grants 20511102200 and 21511102200 , and in part by the Key R&D Program of Guangdong Province under Grant 2020B0101090001.
这项工作得到了国家重点研发计划项目2019 YFB 2101703的支持，海南大学基金项目KYQD 22168的支持，海南省互联网信息检索重点实验室项目2022 KY 01的支持，部分由国家自然科学基金赠款62272107和U19 A2066资助，部分由上海市科技创新行动计划赠款20511102200和21511102200资助，广东省重点研发计划资助项目2020 B 0101090001。
(Corresponding author: Haibin Kan).
（通讯作者：阚海滨）。
Digital Object Identifier no. 10.1109/TSC.2023.3234886
数字对象标识符编号10.1109/TSC.2023.3234886
servers nearby users, which brings convenience in using computational resources, storage resources or bandwidth.
服务器在用户附近，这给用户使用计算资源、存储资源或带宽带来了方便。
Self-driving cars or automated guided vehicles (AGV) are cutting-edge research areas, dedicating to the elimination or reduction of human drivers so that the whole social productivity can be improved significantly. One of the most challenging problems in self-driving is how to avoid collisions. AGV is a typical IoT device, where frequent computation and low bandwidth are required. Previous literature [9], [10], [26], [28], [29] employ fog node or the edge to undertake the computation task. However, little of them considers communication round-efficiency, privacy and anonymity simultaneously. In this paper, we fill the gap by investigating an AGV collision-resistance (or conflict-free) paradigm [9], [10] leveraging edge computing. Further, by listing all occasions that two AGVs meet at a crossroad or an intersection, we design a mechanism to judge whether they may collide with each other.
自动驾驶汽车或自动导引车（AGV）是前沿研究领域，致力于消除或减少人类驾驶员，从而显著提高整个社会的生产力。自动驾驶最具挑战性的问题之一是如何避免碰撞。AGV是典型的物联网设备，需要频繁的计算和低带宽。以前的文献[9]、[10]、[26]、[28]、[29]采用雾节点或边来承担计算任务。然而，很少有人同时考虑通信轮效率、隐私和匿名性。在本文中，我们通过研究利用边缘计算的AGV抗碰撞（或无冲突）范例[9]，[10]来填补差距。此外，通过列出两个AGV在十字路口或交叉口相遇的所有情况，我们设计了一个机制来判断它们是否可能相互碰撞。
In the proposed paradigm, the AGVs, which are moving around, are the central role of the system. An AGV broadcasts information to nearby edge servers, which handle the information and send commands back to the AGV. That means AGVs connect to the edge servers passively and no handshake communication round is required. Edge servers are responsible for generating “virtual” traffic lights for each AGV, ordering it to pass through or stop at a crossroad. Traffic lights are important methods to solve collision problems in signal-free intersection [11], [12], [13].
在所提出的范例中，AGV，这是移动，是系统的核心作用。AGV将信息广播到附近的边缘服务器，边缘服务器处理信息并将命令发送回AGV。这意味着AGV被动地连接到边缘服务器，并且不需要握手通信回合。边缘服务器负责为每个AGV生成“虚拟”交通信号灯，命令它通过或停在十字路口。交通信号灯是解决无信号交叉口碰撞问题的重要手段[11]、[12]、[13]。
However, the privacy of an AGV may leak in such a broadcasting environment. For example, an AGV from one company may collect another company’s AGV information when delivering a package in a residential area. As is known to all, sensitive customer information is critical in business competition. Due to the privacy-leakage problem, we take advantage of a location-based encryption (LBE)
然而，AGV的隐私可能在这样的广播环境中泄漏。例如，当在住宅区递送包裹时，来自一家公司的AGV可能收集另一家公司的AGV信息。众所周知，敏感的客户信息在商业竞争中至关重要。由于隐私泄漏的问题，我们利用基于位置的加密（LBE）
algorithm to protect the AGV’s information. The LBE algorithm is adapted from ciphertext-policy attribute-based encryption (CP-ABE) [14]. This is reasonable and achievable, for we use location coordinates as attribute values. In addition, the attributes in the access policy of CP-ABE enable integer comparison, making data owners encrypt messages with attribute values in an arbitrary scope. Thus, data users with an attribute value within the scope could decrypt the corresponding ciphertexts.
算法来保护AGV的信息。LBE算法改编自密文策略基于属性的加密（CP-ABE）[14]。这是合理的和可实现的，因为我们使用位置坐标作为属性值。此外，CP-ABE访问策略中的属性支持整数比较，使得数据所有者可以在任意范围内使用属性值加密消息。因此，具有范围内的属性值的数据用户可以解密相应的密文。
In our scenario, AGVs encrypt their running state information using their location coordinates and CP-ABE master public key, meeting the requirement that no edge server is involved. Apart from privacy protection, we also benefit from CP-ABE thanks to its ability of one-to-many encryption (or broadcast encryption) [15], [16]. Consequently, only necessary edge servers could decrypt messages and then they send the traffic light results to each AGV. Furthermore, no handshake process is required between AGVs and edge servers.
在我们的场景中，AGV使用其位置坐标和CP-ABE主公钥加密其运行状态信息，满足不涉及边缘服务器的要求。除了隐私保护，我们还受益于CP-ABE，这要归功于它的一对多加密（或广播加密）能力[15]，[16]。因此，只有必要的边缘服务器可以解密消息，然后它们将交通灯结果发送到每个AGV。此外，AGV和边缘服务器之间不需要握手过程。
The contributions of our work are fourfold:
我们工作的贡献有四个方面：
  • We list all occasions that two AGVs may collide in the proposed road model. Then, we resolve the collision problem by introducing a virtual traffic light mechanism.
    我们列出了所有的情况下，两个AGV可能发生碰撞，在建议的道路模型。然后，我们通过引入虚拟交通灯机制来解决碰撞问题。
  • We incorporate a cloud-edge computing architecture for the proposed AGV-centered collision-resistance paradigm. In this paradigm, only edge servers are involved in controlling the AGVs.
    我们将一个云边缘计算架构的建议AGV为中心的碰撞阻力范例。在这种模式中，只有边缘服务器参与控制AGV。
  • To protect the AGVs’ information, we leverage an LBE algorithm based on the CP-ABE access policy. To the best of our knowledge, we are the first to implement LBE by taking advantage of the CP-ABE access policy.
    为了保护AGV的信息，我们利用基于CP-ABE访问策略的LBE算法。据我们所知，我们是第一个利用CP-ABE访问策略实现LBE的公司。
  • Consequently, we achieve favorable properties in privacy-preserving communication, such as noninteractivity and anonymity, fitting well with edge computing architecture. Moreover, we conduct concrete experiments to evaluate the performance of the proposed paradigm.
    因此，我们在隐私保护通信中实现了良好的特性，如非交互性和匿名性，非常适合边缘计算架构。此外，我们进行具体的实验，以评估所提出的范例的性能。

2 Key Techniques  2关键技术

2.1 Edge Computing  2.1边缘计算

Edge computing is a new computing paradigm, differing from traditional cloud computing. Its core idea is to make computing closer to the source of the data [8]. In other words, edge computing is a novel computing model, which unifies the resources close to users in geographical distance or network distance, and provides computing, storage and network services for applications [18]. In the model, cloud servers are only responsible for managing the edge servers and they do not participate in business related to the edge applications. Edge computing is a promising technique in a typical future generation IoT scenario, where IoT devices are heterogeneous and ubiquitous.
边缘计算是一种新的计算模式，不同于传统的云计算。其核心思想是使计算更接近数据的来源[8]。换句话说，边缘计算是一种新型的计算模式，它将地理距离或网络距离上接近用户的资源统一起来，为应用提供计算、存储和网络服务[18]。在该模型中，云服务器只负责管理边缘服务器，不参与与边缘应用相关的业务。边缘计算在典型的下一代物联网场景中是一种很有前途的技术，其中物联网设备是异构的和无处不在的。

2.2 Ciphertext-Policy Attribute-Based Encryption
2.2基于密文策略的属性加密

Ciphertext-policy Attribute-based encryption (CP-ABE) [14], [17] scheme embeds an access policy (AP) in its ciphertext. An AP is composed of attributes, which are written in a
基于属性的加密（CP-ABE）[14]，[17]方案在其密文中嵌入访问策略（AP）。AP由属性组成，这些属性被写入
boolean or threshold-gate expression. The CP-ABE scheme enables a user to decrypt a ciphertext only when its attributes satisfy the ciphertext’s AP. Thus, users encrypt messages using attributes and there might be multiple decryptors for a ciphertext in CP-ABE. The CP-ABE scheme can be formalized with four algorithms, defined as below:
布尔或阈值门表达式。CP-ABE方案使得用户仅当密文的属性满足密文的AP时才能够解密密文。因此，用户使用属性加密消息，并且在CP-ABE中可能存在用于密文的多个解密器。CP-ABE方案可以用四种算法形式化，定义如下：
  • ( M S K , M P K ) ← ( M S K , M P K ) ← (MSK,MPK)larr(M S K, M P K) \leftarrow Setup ABE ( λ ) : ABE  ( λ ) : _("ABE ")(lambda):_{\text {ABE }}(\lambda): The setup algorithm takes in the security parameter λ λ lambda\lambda. It outputs the master key pair (MPK, MSK). The authority holds MSK privately and publishes MPK.
    ( M S K , M P K ) ← ( M S K , M P K ) ← (MSK,MPK)larr(M S K, M P K) \leftarrow 设置 ABE ( λ ) : ABE  ( λ ) : _("ABE ")(lambda):_{\text {ABE }}(\lambda): 设置算法接受安全参数 λ λ lambda\lambda 。它输出主密钥对（MPK，MSK）。管理局私下持有MSK并出版MPK。
  • S K ← KeyGen ABE ( MSK , S ) : S K ← KeyGen ABE ( MSK , S ) : quad SK larrKeyGen_(ABE)(MSK,S):\quad S K \leftarrow \operatorname{KeyGen}_{\mathrm{ABE}}(\mathrm{MSK}, S): The key generation algorithm takes as input the master private key MSK and an attributes set S S SS that describes the key. It outputs a decryption key S K S K SKS K.
    S K ← KeyGen ABE ( MSK , S ) : S K ← KeyGen ABE ( MSK , S ) : quad SK larrKeyGen_(ABE)(MSK,S):\quad S K \leftarrow \operatorname{KeyGen}_{\mathrm{ABE}}(\mathrm{MSK}, S): 密钥生成算法将主私钥MSK和描述密钥的属性集 S S SS 作为输入。它输出解密密钥 S K S K SKS K 。
  • C T ← Encrypt ′ A B E ( M , a p , MPK ) : C T ← Encrypt ′ A B E ( M , a p , MPK ) : CT larrEncrypt^(')_(ABE)(M,ap,MPK):C T \leftarrow \operatorname{Encrypt}^{\prime}{ }_{A B E}(M, a p, \mathrm{MPK}): The encryption algorithm takes in the master public key MPK, a message M ∈ G T M ∈ G T M inG_(T)M \in \mathcal{G}_{T}, and an access policy ap. The algorithm encrypts M M MM and produces a ciphertext C T C T CTC T such that only a user with a set of attributes satisfying the access policy will be able to decrypt the message.
    C T ← Encrypt ′ A B E ( M , a p , MPK ) : C T ← Encrypt ′ A B E ( M , a p , MPK ) : CT larrEncrypt^(')_(ABE)(M,ap,MPK):C T \leftarrow \operatorname{Encrypt}^{\prime}{ }_{A B E}(M, a p, \mathrm{MPK}): 加密算法接受主公钥MPK、消息 M ∈ G T M ∈ G T M inG_(T)M \in \mathcal{G}_{T} 和访问策略ap。该算法加密 M M MM 并产生密文 C T C T CTC T ，使得只有具有满足访问策略的一组属性的用户将能够解密消息。
  • M ← M ← M larrM \leftarrow Decrypt ′ A B E ( C T , S K ) ′ A B E ( C T , S K ) ^(')_(ABE)(CT,SK)^{\prime}{ }_{A B E}(C T, S K) : The decryption algorithm takes as input the master public key MPK, a ciphertext C T C T CTC T (containing an access policy a p a p apa p ) and a decryption key S K S K SKS K (generated from an attributes set S S SS ). If S S SS satisfies a p a p apa p, the algorithm will decrypt the ciphertext and return a message M M MM. Otherwise, it will output ⊥ ⊥ _|_\perp.
    M ← M ← M larrM \leftarrow 解密 ′ A B E ( C T , S K ) ′ A B E ( C T , S K ) ^(')_(ABE)(CT,SK)^{\prime}{ }_{A B E}(C T, S K) ：解密算法将主公钥MPK、密文 C T C T CTC T （包含访问策略 a p a p apa p ）和解密密钥 S K S K SKS K （从属性集 S S SS 生成）作为输入。如果 S S SS 满足 a p a p apa p ，则算法将解密密文并返回消息 M M MM 。否则，它将输出 ⊥ ⊥ _|_\perp 。
    To emphasize, the attributes in CP-ABE access policy supports integer comparison. For example, suppose an access policy a p = a p = ap=a p= "David AND age ≥ 11 ≥ 11 >= 11\geq 11 ". Then an attribute set S 1 = S 1 = S_(1)=S_{1}= [“David”, “age=12”] will satisfy ap, while attribute set S 2 = S 2 = S_(2)=S_{2}= [“David”, “age=10”] will not. The CP-ABE scheme used in our paradigm is inherited from Bethencourt, Sahai and Waters (BSW) implementation [14].
    需要强调的是，CP-ABE访问策略中的属性支持整数比较。例如，假设访问策略 a p = a p = ap=a p= “大卫和年龄 ≥ 11 ≥ 11 >= 11\geq 11 “。那么属性集合 S 1 = S 1 = S_(1)=S_{1}= [“大卫”，“年龄=12”]将满足ap，而属性集合 S 2 = S 2 = S_(2)=S_{2}= [“大卫”，“年龄=10”]将不满足ap。在我们的范例中使用的CP-ABE方案继承自Bethencourt、Sahai和沃茨（BSW）实现[14]。

2.3 Combining AES Encryption With BSW CP-ABE
2.3结合AES加密和BSW CP-ABE

AES [19], short for advanced encryption standard, is a symmetric encryption scheme, which could protect information of arbitrary bytes. The AES scheme roughly has the following two algorithms:
AES [19]是高级加密标准的缩写，是一种对称加密方案，可以保护任意字节的信息。AES方案大致有以下两种算法：
  • c t ← AESEnc ( s e e d , m s g ) : c t ← AESEnc ( s e e d , m s g ) : quad ct larr AESEnc(seed,msg):\quad c t \leftarrow \operatorname{AESEnc}(s e e d, m s g): The AES encryption algorithm takes in a seed for generating the symmetric key, and a plaintext m s g m s g msgm s g. It outputs the ciphertext c t c t ctc t.
    c t ← AESEnc ( s e e d , m s g ) : c t ← AESEnc ( s e e d , m s g ) : quad ct larr AESEnc(seed,msg):\quad c t \leftarrow \operatorname{AESEnc}(s e e d, m s g): AES加密算法接受用于生成对称密钥的种子和明文 m s g m s g msgm s g 。它输出密文 c t c t ctc t 。
  • m s g ← AESDec ( s e e d , c t ) : m s g ← AESDec ( s e e d , c t ) : msg larr AESDec(seed,ct):m s g \leftarrow \operatorname{AESDec}(s e e d, c t): The AES decryption algorithm takes as input a seed and a ciphertext c t c t ctc t. It outputs a plaintext msg .
    m s g ← AESDec ( s e e d , c t ) : m s g ← AESDec ( s e e d , c t ) : msg larr AESDec(seed,ct):m s g \leftarrow \operatorname{AESDec}(s e e d, c t): AES解密算法将种子和密文 c t c t ctc t 作为输入。它输出一个明文msg。
    As defined in Section 2.2, CP-ABE is extended from public-key encryption scheme and it only allows one to encrypt a message on group G T G T G_(T)\mathcal{G}_{T}. To enable encrypting an arbitrary string m s g m s g msgm s g, we combine AES encryption with CP-ABE scheme. To simplify the writing, we integrate AES encryption into the CP-ABE scheme. Thus, we redefine the CP-ABE encryption and decryption algorithms as Fig. 1 shows.
    如第2.2节所定义，CP-ABE是从公钥加密方案扩展而来的，它只允许加密组 G T G T G_(T)\mathcal{G}_{T} 上的消息。为了能够加密任意字符串 m s g m s g msgm s g ，我们将联合收割机AES加密与CP-ABE方案相结合。为了简化编写，我们将AES加密集成到CP-ABE方案中。因此，我们重新定义了CP-ABE加密和解密算法，如图1所示。

3 Problem Statement  3问题陈述

We aim at establishing an edge-cloud-based architecture to achieve a collision-resistance paradigm for AGVs. To protect the privacy of the broadcast messages, we incorporate a
我们的目标是建立一个基于边缘云的架构，以实现一个抗碰撞的AGV范例。为保障广播讯息的私隐，我们在
\((c t, C T) \leftarrow \operatorname{Encrypt}_{\mathrm{ABE}}(m s g, a p, \mathrm{MPK}):\)
    \(M \stackrel{R}{\leftarrow} \mathcal{G}_{T}\)
    \(C T \leftarrow \operatorname{Encrypt}_{\mathrm{ABE}}^{\prime}(M, a p, \mathrm{MPK})\)
    \(c t \leftarrow\) AESEnc (M.toBytes(), msg)
    return (ct, \(C T\) )
\(\underline{m s g} \leftarrow \operatorname{Decrypt}_{\mathrm{ABE}}(C T, c t, S K):\)
    \(M \leftarrow \operatorname{Decrypt}_{\mathrm{ABE}}^{\prime}(C T, S K)\)
    \(m s g \leftarrow \operatorname{AESDec}(\) M.toBytes ()\(, c t)\)
    return \(m s g\)
Fig. 1. Encryption of arbitrary message using CP-ABE and the corresponding decryption algorithm.
图1.使用CP-ABE对任意消息进行加密及相应的解密算法。
LBE scheme where AGVs’ positions are used when encryption. The LBE scheme is implemented based on the CP-ABE algorithm. We assume a global key pair (GPK, GSK) which is shared among cloud server and edge servers. Moreover, it is updated periodly, e.g., daily or weekly. The global public key GPK is used protect privacy of access policy by AGVs. We choose ElGamal as the underlying scheme and omit the details on how to encrypt access policies in this paper.
加密时使用AGV位置的LBE方案。LBE方案是基于CP-ABE算法实现的。我们假设在云服务器和边缘服务器之间共享的全局密钥对（GPK，GSK）。此外，它定期更新，例如，每天或每周。全局公钥GPK用于保护AGV访问策略的隐私。我们选择ElGamal作为底层方案，并省略了本文中如何加密访问策略的细节。
We take a wharf yard as an example to untangle how our proposed paradigm avoids collision at the crossroads. AGVs perform tasks by carrying specific materials from one place to another, under the direction of a path planning system, on the yard. The path planning system is out of the discussion and we only make use of its outputs. Multiple AGVs may arrive at a crossroad at the same time. AGVs may crash if given no proper coordination program, leading to traffic disasters. The collision-resistance paradigm should guarantee the safety of each AGV and the carrying capacity of the yard simultaneously.
我们以一个码头堆场为例，来解开我们所提出的范例如何避免在十字路口发生碰撞。AGV通过在路径规划系统的指导下将特定材料从一个地方运送到另一个地方来执行任务。路径规划系统不在讨论范围内，我们只利用其输出。多个AGV可能同时到达十字路口。如果没有适当的协调程序，AGV可能会崩溃，导致交通灾难。防碰撞范例应同时保证每个AGV的安全和堆场的承载能力。
All AGVs are authenticated and they are equipped with different kinds of reliable sensors. GPS module enables an AGV to locate its coordinate accurately, and the networking module allows the AGV to deliver its location information every hundreds of milliseconds to edge servers. An AGV is roughly about 10 meters long. The speed also differs from another with a different model, weight capacity or task priority. The driving direction of an AGV is either forwarding, left or right. If an AGV turns left/right, it should be on the left/right lane. We assume that there are two lanes in each direction without loss of generality. We do not consider insider attack, replay attack or DoS attack of corrupted AGVs in this paper. 1 1 ^(1){ }^{1}
所有AGV都经过认证，并配备了不同类型的可靠传感器。GPS模块使AGV能够准确定位其坐标，网络模块允许AGV每隔数百毫秒将其位置信息传递到边缘服务器。AGV大约有10米长。速度也不同于另一个具有不同的模型，重量容量或任务优先级。AGV的行驶方向是向前、向左或向右。如果AGV左/右转弯，它应该在左/右车道上。我们假设在每个方向上有两个车道而不失一般性。我们没有考虑内部攻击，重放攻击或拒绝服务攻击的损坏AGV在本文中。 1 1 ^(1){ }^{1}
To simplify the road map of the yard, we only consider a crossroad with four directions, i.e., east, west, south, north. East (or west) to west (or east) is the latitude direction and south (or north) to north (or east) is the longitude direction. We denote α α alpha\alpha as the heading direction. Thus, α α alpha\alpha can only be 0 , 90,180 or 270 . Also, a virtual point is defined every 10 meters along both the latitude direction and longitude direction. We use such points as the location coordinates of
为了简化庭院的路线图，我们只考虑一个四个方向的十字路口，即，东西南北东（或西）到西（或东）是纬度方向，南（或北）到北（或东）是经度方向。我们将 α α alpha\alpha 表示为前进方向。因此， α α alpha\alpha 只能是0、90、180或270。此外，沿着纬度方向和经度方向每隔沿着10米定义一个虚拟点。我们使用这些点作为
  1. If an AGV shows anomalies, rather than being corrupted, the replay problem can be resolved with a timestamp attached within broadcasted messages; the DoS problem can be detected with other monitoring modules.
    如果AGV显示异常，而不是被损坏，则可以通过在广播消息中附加时间戳来解决重放问题;可以通过其他监控模块检测DoS问题。
Fig. 2. An example containing edge servers, crossroads and an AGV
图2.包含边缘服务器、十字路口和AGV的示例
AGVs. Thus, if the yard is 3 km ∗ 3 km 3 km ∗ 3 km 3km**3km3 \mathrm{~km} * 3 \mathrm{~km}, then there are 300 ∗ 300 300 ∗ 300 300**300300 * 300 points on the map. We name the points as ( 1 , 1 ) ( 1 , 1 ) (1,1)(1,1) to ( 300 , 300 ) ( 300 , 300 ) (300,300)(300,300) from the top left corner to the bottom right corner of the yard, respectively.
AGV。因此，如果码是 3 km ∗ 3 km 3 km ∗ 3 km 3km**3km3 \mathrm{~km} * 3 \mathrm{~km} ，则地图上有 300 ∗ 300 300 ∗ 300 300**300300 * 300 点。从院子的左上角到右下角，我们分别将这些点命名为 ( 1 , 1 ) ( 1 , 1 ) (1,1)(1,1) 到 ( 300 , 300 ) ( 300 , 300 ) (300,300)(300,300) 。
To maximize the transportation efficiency on the yard, we work out a new traffic light rule-each AGV has a separated traffic light result at each crossroad, rather than a unified light result for all the AGVs in one direction. This is reasonable, since the distance between two neighboring crossroads might be short and AGVs have no time to slow down if given a unified light result. An edge server is responsible for generating the signals of traffic lights for one or multiple crossroads.
为了最大限度地提高堆场的运输效率，我们制定了一个新的交通灯规则-每个AGV在每个十字路口都有一个单独的交通灯结果，而不是一个统一的交通灯结果。这是合理的，因为两个相邻十字路口之间的距离可能很短，如果给出统一的灯光结果，AGV没有时间减速。边缘服务器负责生成一个或多个十字路口的交通灯信号。
Fig. 2 shows an example that an AGV is at ( x , y ) ( x , y ) (x,y)(x, y), heading north (i.e., α = 90 α = 90 alpha=90\alpha=90 ). All edge servers in the blue circle can receive encrypted messages broadcast by the AGV. As desired, only the edge servers in the gray rectangle are able to decrypt the messages. Finally, edge servers will produce traffic lights for AGVs if they get AGVs’ original messages.
图图2示出了AGV位于 ( x , y ) ( x , y ) (x,y)(x, y) ，向北行驶（即， α = 90 α = 90 alpha=90\alpha=90 ）。蓝色圆圈中的所有边缘服务器都可以接收AGV广播的加密消息。根据需要，只有灰色矩形中的边缘服务器能够解密消息。最后，如果边缘服务器获得AGV的原始消息，它们将为AGV生成交通信号灯。

4 LBE FROM BSW CP-ABE
4 LBE从BSW CP-ABE

4.1 Construction  4.1建设

Location-based encryption (LBE) [20], [21] is a cryptographic scheme aiming at securing mobile communication by limiting the area inside which the intended recipient can decrypt a message. Formally, LBE can be defined as below, inherited from definition [21]:
基于位置的加密（LBE）[20]，[21]是一种加密方案，旨在通过限制预期接收者可以解密消息的区域来保护移动的通信。形式上，LBE可以定义如下，继承自定义[21]：
  • Setup ( λ ) Setup ( λ ) Setup(lambda)\operatorname{Setup}(\lambda) : The setup algorithm takes in the security parameter λ λ lambda\lambda. It outputs the master key pair (MPK, MSK).
    Setup ( λ ) Setup ( λ ) Setup(lambda)\operatorname{Setup}(\lambda) ：设置算法接受安全参数 λ λ lambda\lambda 。它输出主密钥对（MPK，MSK）。
  • KeyGen(MSK, [ x , y ] ) [ x , y ] ) [x,y])[x, y]) : The key generation algorithm takes as input the master private key MSK and a location ( x , y ) ( x , y ) (x,y)(x, y), where x , y ∈ Z x , y ∈ Z x,y inZx, y \in \mathcal{Z}. It outputs a decryption key S K S K SKS K.
    KeyGen（MSK， [ x , y ] ) [ x , y ] ) [x,y])[x, y]) ）：密钥生成算法将主私钥MSK和位置 ( x , y ) ( x , y ) (x,y)(x, y) 作为输入，其中 x , y ∈ Z x , y ∈ Z x,y inZx, y \in \mathcal{Z} 。它输出解密密钥 S K S K SKS K 。
  • Encrypt( m s g , [ x 1 , y 1 , x 2 , y 2 ] m s g , x 1 , y 1 , x 2 , y 2 msg,[x_(1),y_(1),x_(2),y_(2)]m s g,\left[x_{1}, y_{1}, x_{2}, y_{2}\right], MPK) : The encryption algorithm takes in the master public key MPK, a message m s g m s g msgm s g, and two geographical point ( x 1 , y 1 ) x 1 , y 1 (x_(1),y_(1))\left(x_{1}, y_{1}\right) and ( x 2 , y 2 ) x 2 , y 2 (x_(2),y_(2))\left(x_{2}, y_{2}\right), where x 1 , y 1 , x 2 , y 2 ∈ Z x 1 , y 1 , x 2 , y 2 ∈ Z x_(1),y_(1),x_(2),y_(2)inZx_{1}, y_{1}, x_{2}, y_{2} \in \mathcal{Z}. The algorithm produces the ciphertext ( C T , c t C T , c t CT,ctC T, c t ).
    加密（ m s g , [ x 1 , y 1 , x 2 , y 2 ] m s g , x 1 , y 1 , x 2 , y 2 msg,[x_(1),y_(1),x_(2),y_(2)]m s g,\left[x_{1}, y_{1}, x_{2}, y_{2}\right] ，MPK）：加密算法接受主公钥MPK、消息 m s g m s g msgm s g 和两个地理点 ( x 1 , y 1 ) x 1 , y 1 (x_(1),y_(1))\left(x_{1}, y_{1}\right) 和 ( x 2 , y 2 ) x 2 , y 2 (x_(2),y_(2))\left(x_{2}, y_{2}\right) ，其中 x 1 , y 1 , x 2 , y 2 ∈ Z x 1 , y 1 , x 2 , y 2 ∈ Z x_(1),y_(1),x_(2),y_(2)inZx_{1}, y_{1}, x_{2}, y_{2} \in \mathcal{Z} 。算法产生密文（ C T , c t C T , c t CT,ctC T, c t ）。
  • Decrypt ( C T , c t , S K ) Decrypt ( C T , c t , S K ) Decrypt(CT,ct,SK)\operatorname{Decrypt}(C T, c t, S K) : The decryption algorithm takes as input the master public key MPK, a ciphertext ( C T , c t ) ( C T , c t ) (CT,ct)(C T, c t), and a decryption key S K S K SKS K. Only if x 1 < x < x 1 < x < x_(1) < x <x_{1}<x< x 2 x 2 x_(2)x_{2} and x 2 < y < y 2 x 2 < y < y 2 x_(2) < y < y_(2)x_{2}<y<y_{2}, the algorithm will decrypt the
    Decrypt ( C T , c t , S K ) Decrypt ( C T , c t , S K ) Decrypt(CT,ct,SK)\operatorname{Decrypt}(C T, c t, S K) ：解密算法将主公钥MPK、密文 ( C T , c t ) ( C T , c t ) (CT,ct)(C T, c t) 和解密密钥 S K S K SKS K 作为输入。只有当 x 1 < x < x 1 < x < x_(1) < x <x_{1}<x< x 2 x 2 x_(2)x_{2} 和 x 2 < y < y 2 x 2 < y < y 2 x_(2) < y < y_(2)x_{2}<y<y_{2} 时，算法才会解密