Urban air mobility: A comprehensive review and comparative analysis with autonomous and electric ground transportation for informing future research
城市空中移动：一项综合评估和与自动驾驶和电动地面交通的比较分析，以指导未来研究

2.1. History of UAM service and value estimates for the emerging eVTOL UAM market
2.1. UAM 服务的历史和新兴 eVTOL UAM 市场的价值估计

The concept of urban air mobility is not new, with examples of UAM services using helicopters dating to the 1940s. From 1947 to 1971, Los Angeles Airways used helicopters to transport people and mail in the Los Angeles area, including between Disneyland and the Los Angeles International Airport (LAX). Los Angeles Airways experienced two accidents caused by mechanical failure in 1968 and subsequently ceased operations (Harrison, 2017, as referenced in Thipphavong et al., 2018). From 1953 to 1979, New York Airways used helicopters to fly passengers between Manhattan locations and the three major airports in New York City (Newark Liberty International Airport [EWR], LaGuardia Airport [LGA], and John F. Kennedy International Airport [JFK]). This service similarly ceased due to several accidents caused by mechanical failure (Witken, 1979, as referenced in Thipphavong et al., 2018, Mayor and Anderson, 2019). The cost of a passenger ticket on a New York Airways shuttle was between $5 and $9, or about $47 to $86 in 2019 dollars (Mayor and Anderson, 2019). These early UAM operations successfully operated for more than two decades, ultimately ceasing operations due to safety concerns. These historic examples provide evidence of the potential value to consumers for similar (albeit safer) UAM services today.
城市空中移动的概念并不新鲜，早在 1940 年代就有使用直升机的 UAM 服务的例子。从 1947 年到 1971 年，洛杉矶航空公司在洛杉矶地区使用直升机运送人员和邮件，包括迪士尼乐园和洛杉矶国际机场（LAX）之间的运输。洛杉矶航空公司在 1968 年发生了两起由机械故障引起的事故，随后停止运营（Harrison, 2017，参考文献为 Thipphavong 等，2018）。从 1953 年到 1979 年，纽约航空公司使用直升机在曼哈顿各地和纽约市的三个主要机场（纽瓦克自由国际机场[EWR]，拉瓜迪亚机场[LGA]和约翰·肯尼迪国际机场[JFK]）之间运送乘客。由于多起机械故障引起的事故，该服务也停止运营（Witken, 1979，参考文献为 Thipphavong 等，2018，Mayor 和 Anderson，2019）。纽约航空公司的乘客票价在 5 美元到 9 美元之间，相当于 2019 年的 47 美元到 86 美元（Mayor 和 Anderson，2019）。这些早期的 UAM 运营成功运营了二十多年，最终因安全问题而停止运营。 这些历史例子为今天类似（尽管更安全）的城市空中交通服务对消费者的潜在价值提供了证据。

Several helicopter operators are providing on-demand urban passenger air service. BLADE operates³ between various locations in Manhattan and one of the three main airports in New York City (JFK, LGA, and EWR). Flights are bookable within 30 min of departure and the one-way cost is $195; additional charges starting at $85 apply for baggage above 25 lb (that is transported via a ground service), last-minute bookings, and cancellations received within three hours of departure (BLADE, 2020). In 2019, Uber partnered with HeliFlite⁴ to offer flights from Manhattan to JFK airport between the hours of 1 PM and 6 PM, Monday through Friday. Uber Copter can be booked on demand or up to five days before the flight (Matthews, 2019). The helicopter option appears on the Uber app and the one-way cost ranges from $200 to $225 per person and includes one piece of luggage up to 50 lb (Ballentine, 2019, Uber, 2020). In 2016, Airbus started Voom, an on-demand booking platform that connected travelers to helicopter service providers in São Paulo, Brazil, and later expanded service to Mexico City and the San Francisco Bay Area before permanently ceasing operations due to COVID-19 in April of 2020 (Airbus, 2020b). Flights in Mexico City were bookable within 60 min of departure of the flight or could be reserved up to seven days in advance (Airbus, 2018).
几家直升机运营商提供按需的城市客运航空服务。BLADE 在曼哈顿的各个地点和纽约市的三个主要机场（JFK、LGA 和 EWR）之间运营航班。航班可在起飞前 30 分钟内预订，单程费用为 195 美元；超过 25 磅的行李（通过地面服务运输）将收取额外费用，临时预订和起飞前三小时内取消预订也将收取额外费用（BLADE，2020 年）。2019 年，优步与 HeliFlite 合作，提供从曼哈顿到 JFK 机场的航班，时间为周一至周五的下午 1 点至下午 6 点。优步直升机可以按需预订，或在飞行前最多提前五天预订（Matthews，2019 年）。直升机选项出现在优步应用程序上，单程费用为每人 200 至 225 美元，包括一件重量不超过 50 磅的行李（Ballentine，2019 年，Uber，2020 年）。 2016 年，空中客车公司推出了 Voom，一个按需预订平台，将旅客与圣保罗、巴西的直升机服务提供商连接起来，并在之后扩展到墨西哥城和旧金山湾区，但由于 COVID-19 的影响，该平台于 2020 年 4 月永久停止运营（Airbus，2020b）。在墨西哥城，航班可以在起飞前 60 分钟内预订，或者提前最多七天预订（Airbus，2018）。

These modern-day on-demand helicopter services have been important to UAM researchers, as they provided information about customer preferences (booking patterns, willingness to pay, most popular routes) and “operational challenges related to a lack of infrastructure, public acceptance, [and] on-demand versus scheduled routes” (Airbus, 2020b). They help set the context for how the UAM community is envisioning the possibilities for stimulating demand through lower per-passenger mile (pax-mile) operating costs with new eVTOL aircraft.
这些现代化的按需直升机服务对城市空中交通研究人员来说非常重要，因为它们提供了有关客户偏好（预订模式、支付意愿、最受欢迎的航线）以及“基础设施缺乏、公众接受度、按需与定期航线之间的运营挑战”（Airbus，2020b）的信息。它们帮助我们了解城市空中交通社区如何通过新的电动垂直起降飞行器降低每乘客英里（pax-mile）运营成本来刺激需求的可能性。

BLADE and Uber Copter charge about $30 per pax-mile in Manhattan, whereas Voom charged about $10 per pax-mile (Booz Allen Hamilton, 2018, Reiche et al., 2018). Uber Elevate estimates the cost of a passenger helicopter service at about $8.93 per pax-mile (Holden, 2018), and McKinsey and Company estimates the cost between $6 and $8 per seat-mile⁵ (Johnson, Riedel, and Sahdev, 2020). Uber Elevate has reported that they anticipate at the launch of their on-demand air taxi that service costs will be $5.73 per pax-mile but will decrease in the near term to $1.84 by increasing utilization through ride-hailing (Holden, 2018). In later conferences, Uber Elevate noted that at $2.00 per pax-mile, the flight operating cost would be $662/hr as compared to $1,253/hr that is more common among helicopters operating today (Uber Elevate, 2019). On an hourly basis, longer-term, Uber Elevate anticipates that advancements in manufacturing and autonomy will decrease both fixed and variable costs, resulting in $0.44 per pax-mile cost; in comparison, in 2017, the American Automobile Association (AAA) estimated the full cost of auto ownership in the U.S. to be between $0.46 and $0.61 per mile (Holden, 2018, AAA, 2017). The McKinsey report estimates near-term costs between $2.50 and $4.50 per seat-mile and long-term costs between $0.50 and $2.50 (Johnson, Riedel, and Sahdev, 2020). The UAM cost estimates provided by Uber Elevate and McKinsey and Company are optimistic compared to other reports, such as one conducted for the National Aeronautics and Space Administration (NASA) that forecasts the costs of a five-seat eVTOL at $6.25/pax-mile in the near term but “in the long term, operational efficiency, autonomy, technology improvements may decrease costs by 60%” (i.e., $3.75/pax-mile) (Booz Allen Hamilton, 2018, Reiche et al., 2018). Although the cost estimates of providing UAM service vary, near-term and long-term eVTOL operations will likely operate at lower costs compared to current helicopter service, resulting in more demand for UAM service. In comparison, costs for shared autonomous vehicles in the 2030s to 2040s have been estimated by multiple studies to range from $0.29 to $0.49 per-seat-mile (Johnson, 2015, Albright et al., 2015, Chen et al., 2016, Corwin et al., 2016), with lower cost estimates corresponding to purpose-built SAVs. These costs range from five times to ten times lower than the UAM cost estimates noted above in the mid-term to near-parity with UAM in the long-term; however, direct comparisons are challenging at present because of the uncertainty in the technologies and development timelines for both SAVs and UAM.
在曼哈顿，BLADE 和 Uber Copter 每人每英里收费约 30 美元，而 Voom 每人每英里收费约 10 美元（Booz Allen Hamilton，2018 年，Reiche 等，2018 年）。Uber Elevate 估计乘客直升机服务的成本约为每人每英里 8.93 美元（Holden，2018 年），而麦肯锡公司估计每个座位每英里的成本在 6 至 8 美元之间（Johnson，Riedel 和 Sahdev，2020 年）。Uber Elevate 报告称，他们预计在他们的按需空中出租车推出时，服务成本将为每人每英里 5.73 美元，但通过增加乘车率，这一成本将在短期内降至 1.84 美元（Holden，2018 年）。在后来的会议上，Uber Elevate 指出，以每人每英里 2.00 美元的价格，飞行运营成本将为每小时 662 美元，而现今直升机的运营成本通常为每小时 1253 美元（Uber Elevate，2019 年）。从长期来看，Uber Elevate 预计制造业和自主技术的进步将降低固定和可变成本，导致每人每英里成本为 0.44 美元；相比之下，2017 年，美国汽车协会（AAA）估计美国汽车所有权的全部成本在 0.46 至 0.61 美元每英里（Holden，2018 年，AAA，2017 年）。麦肯锡报告估计，短期成本在每个座位英里 2.50 美元至 4.50 美元之间，长期成本在每个座位英里 0.50 美元至 2.50 美元之间（Johnson，Riedel 和 Sahdev，2020 年）。Uber Elevate 和麦肯锡公司提供的城市空中出行成本估计相对于其他报告来说是乐观的，例如美国国家航空航天局（NASA）进行的一项研究预测，短期内五座电动垂直起降飞行器的成本为每人英里 6.25 美元，但“长期来看，运营效率、自主性和技术改进可能会降低成本 60%”（即每人英里 3.75 美元）（Booz Allen Hamilton，2018 年，Reiche 等，2018 年）。尽管提供城市空中出行服务的成本估计各不相同，但短期和长期的电动垂直起降飞行器运营成本可能会比目前的直升机服务更低，从而导致对城市空中出行服务的需求增加。相比之下，2030 年至 2040 年共享自动驾驶车辆的成本估计在每个座位英里 0.29 美元至 0.49 美元之间（Johnson，2015 年，Albright 等，2015 年，Chen 等，2016 年，Corwin 等，2016 年），较低的成本估计对应于专门设计的自动驾驶车辆。 这些成本比上述中期的 UAM 成本估计低五倍到十倍，长期来看与 UAM 成本基本持平；然而，由于 SAV 和 UAM 的技术和发展时间表的不确定性，目前很难进行直接比较。

2.2. Overview of eVTOL aircraft designs
2.2. 电垂直起降飞行器设计概述

Worldwide, there are multiple efforts focused on designing eVTOL aircraft, and more than $2B has been invested in this industry (Sherman, 2020). Collectively, these designs represent fundamentally different design concepts. Multiple publications provide overviews of the different technical specifications and characteristics associated with eVTOL aircraft (e.g., see Roland Berger, 2018, Porsche Consulting, 2018). The Vertical Flight Society (VFS) provides one of the more thorough overviews of the different types of eVTOL aircraft and maintains a database of known eVTOL designs (Electric VTOL News™, 2020). According to VFS, as of March 5, 2020, there were a total of 260 aircraft⁶ that included 99 vectored thrust, 39 lift + cruise, 26 wingless multicopters, 46 hover bikes/flying devices⁷ and 20 eHelos and eGyros (Sherman, 2020). Across these designs, there are large variations in the number of seats, speed, and range.
全球范围内，有多个项目致力于设计电动垂直起降飞行器（eVTOL），已经投资了超过 20 亿美元在这个行业（Sherman，2020）。这些设计代表了根本不同的设计概念。多个出版物提供了关于 eVTOL 飞行器的不同技术规格和特性的概述（例如，参见 Roland Berger，2018 年，Porsche Consulting，2018 年）。垂直飞行学会（VFS）提供了对不同类型的 eVTOL 飞行器的更全面的概述，并维护着已知 eVTOL 设计的数据库（Electric VTOL News™，2020）。根据 VFS 的数据，截至 2020 年 3 月 5 日，共有 260 架飞行器，其中包括 99 架向量推力、39 架升力+巡航、26 架无翼多旋翼、46 架悬停摩托车/飞行装置和 20 架电动直升机和电动陀螺（Sherman，2020）。在这些设计中，座位数、速度和航程存在很大的差异。

Vectored thrust aircraft can use any of their thrusters⁸ for both lift and cruise; representative examples include the Lilium Jet (2 to 5 seats; 186 mph; 186-mile range), Airbus A³ Vahana (1 seat; 118 mph; 31-mile range), Bell Nexus 4EX (5 seats; 150 mph; 150-mile range), and Joby S4 (5 seats; 200 mph; 150-mile range) (Lilium, 2020, Hawkins, 2019, Airbus, 2020a, Bell Flight, 2020, Pope, 2019, Goldstein, 2019, Bogaisky, 2020). According to Sherman, vectored thrust designs—the most common among potential eVTOL designs—will likely be the most efficient eVTOL aircraft but also likely the most difficult to bring to market due to the complexity of designing the aircraft to safety transition between vertical flight and forward flight.
矢量推力飞机可以使用任何一个推力器来实现升力和巡航；代表性的例子包括 Lilium Jet（2 至 5 个座位；186 英里/小时；186 英里航程）、Airbus A Vahana（1 个座位；118 英里/小时；31 英里航程）、Bell Nexus 4EX（5 个座位；150 英里/小时；150 英里航程）和 Joby S4（5 个座位；200 英里/小时；150 英里航程）（Lilium，2020 年，Hawkins，2019 年，Airbus，2020a，Bell Flight，2020 年，Pope，2019 年，Goldstein，2019 年，Bogaisky，2020 年）。根据 Sherman 的说法，矢量推力设计是潜在 eVTOL 设计中最常见的，可能是最高效的 eVTOL 飞机，但由于设计飞机在垂直飞行和前向飞行之间的安全过渡的复杂性，也可能是最难进入市场的。

The lift + cruise is another popular aircraft category under development that has two sets of independent thrusters—one set that is used only for cruise and a second set that is used only for vertical lift. Lift during cruise flight is provided by one or several wings. Representative examples include the Aurora Flight Sciences Pegasus (2 seats; 112 mph; 50-mile range; Aurora Flight Sciences, 2020), EmbraerX Eve⁹ (5 seats; speed and range not public), and Wisk Cora (2 seats; 100 mph; 25-mile range) (Electric VTOL News™, n.d.a, Electric VTOL News™, n.d.b, EmbraerX, 2020, Wisk, 2020).
垂直起降+巡航是另一种正在开发中的流行飞行器类型，它有两套独立的推进器，一套仅用于巡航，另一套仅用于垂直起降。巡航飞行时，提供升力的是一个或多个机翼。代表性的例子包括 Aurora Flight Sciences Pegasus（2 个座位；112 英里/小时；50 英里航程；Aurora Flight Sciences，2020）、EmbraerX Eve（5 个座位；速度和航程未公开）和 Wisk Cora（2 个座位；100 英里/小时；25 英里航程）（Electric VTOL News™，无日期 a，Electric VTOL News™，无日期 b，EmbraerX，2020，Wisk，2020）。

Wingless multicopters are another common design that use their thrusters to produce lift not only for takeoff and vertical flight but for cruise, as well. Representative examples of these aircraft include the Volocopter VC200 (2 seats; 50–62 mph; 19-mile range), the eHang 216 (1 seat; 81 mph; 22-mile range), and the LIFT Aircraft Hexa (1 seat; 60 mph; 12–15 mile range) (Volocopter, 2018, Volocopter, 2020, eHang, 2020a, LIFT Aircraft, 2020). The LIFT Aircraft Hexa is an ultralight passenger air vehicle that seats one passenger who controls the aircraft. Ultralight aircraft will be restricted to recreational use and speeds of under 60 mph but will likely be some of the first eVTOL aircraft to enter the market, as they do not require aircraft and pilot certification under Federal Aviation Regulations Part 103 (FAA, 1982, as noted by Sherman, 2020).
无翼多旋翼机是另一种常见设计，它们利用推进器产生升力，不仅用于起飞和垂直飞行，还用于巡航。这类飞机的代表性例子包括 Volocopter VC200（2 座位；50-62 英里/小时；19 英里航程）、eHang 216（1 座位；81 英里/小时；22 英里航程）和 LIFT Aircraft Hexa（1 座位；60 英里/小时；12-15 英里航程）（Volocopter，2018 年；Volocopter，2020 年；eHang，2020a 年；LIFT Aircraft，2020 年）。LIFT Aircraft Hexa 是一种超轻型乘客飞行器，可容纳一名乘客控制飞行。超轻型飞行器将受到限制，仅限于娱乐用途和时速不超过 60 英里，但很可能成为首批进入市场的电动垂直起降飞行器，因为它们不需要根据联邦航空法规第 103 部分获得飞行器和飞行员认证（FAA，1982 年，如 Sherman，2020 年所述）。

Rotorcraft designs are another area being considered for UAM applications. These concepts include both electric helicopters and novel autogyros (i.e., helicopter-like aircraft in which the rotor rotates not by shaft power from the engine but by the force of air flowing through it; propulsion in forward flight is provided by a separate propeller). Representative examples include the Jaunt Air Mobility gyrocopter (5 seats; 175 mph; range unknown) and the Pal-V Pioneer flying car (2 seats; 99–112 mph; 250–300-mile range) (Jaunt Air Mobility, 2020, Blain, 2020, PAL-V, 2020). Although the popular press often refers to UAM/eVTOL aircraft as “flying cars,” these aircraft typically do not meet the historical definition of a “roadable aircraft” that can be both driven on the ground as a car and flown as an airplane. However, the Pal-V is a roadable aircraft. The Pal-V is not an electric aircraft (hence not an eVTOL) and is intended for personal use and not necessarily UAM; however, it is representative of the novel rotorcraft configurations being explored in the UAM field. Fig. 2 provides examples of each of the aircraft designs discussed above.
旋翼飞行器设计是另一个被考虑用于城市空中交通应用的领域。这些概念包括电动直升机和新颖的自转旋翼机（即，类似直升机的飞行器，其旋翼不是通过发动机的轴动力旋转，而是通过空气流经旋翼产生的力；前进飞行时由独立的螺旋桨提供推进力）。代表性的例子包括 Jaunt Air Mobility 的陀螺旋翼机（5 个座位；175 英里/小时；航程未知）和 Pal-V Pioneer 飞行汽车（2 个座位；99-112 英里/小时；250-300 英里航程）（Jaunt Air Mobility, 2020, Blain, 2020, PAL-V, 2020）。尽管大众媒体经常将城市空中交通/电动垂直起降飞行器称为“飞行汽车”，但这些飞行器通常不符合“可上路飞行器”的历史定义，即既可以作为汽车在地面上行驶，又可以作为飞机飞行。然而，Pal-V 是一种可上路飞行器。Pal-V 不是电动飞行器（因此不是电动垂直起降飞行器），旨在个人使用，而不一定是城市空中交通；然而，它代表了城市空中交通领域正在探索的新颖旋翼飞行器配置。图 2 提供了上述每种飞行器设计的示例。

1. Download : Download high-res image (187KB)
   下载：下载高分辨率图像（187KB）
2. Download : Download full-size image
   下载：下载完整尺寸的图片

Fig. 2. Examples of different UAM aircraft designs.
图 2. 不同的城市空中交通飞行器设计示例。

Image credits: Joby Aviation, 2021, Lilium, 2021, Beta, 2021, Wisk Aero, 2021, Lift, 2021, Jaunt Air Mobility. 2021.
图片来源：Joby Aviation，2021 年，Lilium，2021 年，Beta，2021 年，Wisk Aero，2021 年，Lift，2021 年，Jaunt Air Mobility，2021 年。

What is clear from the discussion above is that there is currently a lack of convergence in designs and underlying business models envisioned by the eVTOL community. The non-convergence of design concepts reflects the novelty of these new battery and electric propulsion technologies and uncertainties regarding how these new technologies will impact aircraft performance. It also reflects a lack of consensus on which missions (or market segments) these aircraft can profitably serve, and whether the aircraft should be flown by the passenger, an onboard pilot, a remote pilot, or autonomously.
从上面的讨论中可以清楚地看出，目前电垂直起降飞行器（eVTOL）社区在设计和基础商业模式上存在着缺乏一致性。设计概念的不一致反映了这些新型电池和电动推进技术的新颖性，以及对这些新技术将如何影响飞机性能的不确定性。这也反映出对这些飞机能够盈利地服务于哪些任务（或市场细分）以及飞机应该由乘客、机上飞行员、远程飞行员还是自主飞行。

3.1. Methodology and scope of review
3.1. 检查的方法和范围

To identify relevant publications in UAM, we conducted a keyword search of “urban air mobility,” “air taxi,” and “UAM” in the AIAA publication database. A similar search was conducted on Scopus using the same keywords but adding exclusion terms for “drone” and “UAV.” The searches were initially conducted in the spring of 2020 and were updated in mid-July 2020.
为了确定与城市空中移动相关的出版物，我们在 AIAA 出版物数据库中进行了关键词搜索，关键词包括“城市空中移动”、“空中出租车”和“UAM”。在 Scopus 上进行了类似的搜索，使用相同的关键词，但添加了排除词“无人机”和“无人机”。这些搜索最初是在 2020 年春季进行的，并在 2020 年 7 月中旬进行了更新。

The search results included journal and conference publications relevant to UAM that were published from January 1, 2015, to June 30, 2020, in which the aforementioned keywords appeared in the title or abstract. A total of 251 publications were identified from the AIAA database and an additional 61 from the Scopus search.
搜索结果包括了从 2015 年 1 月 1 日到 2020 年 6 月 30 日发表的与城市空中交通相关的期刊和会议论文，其中标题或摘要中出现了上述关键词。从 AIAA 数据库中找到了 251 篇论文，从 Scopus 搜索中找到了额外的 61 篇。

To identify relevant EV and AV articles, we reviewed the table of contents of key journals from the transportation field from January 2015 to June 2020¹⁰ and identified articles that were relevant based on their titles and abstracts. We explicitly decided not to use a keyword search for this part of the analysis, so that we could go through the titles and identify publications that were relevant to UAM research, such as ride-hailing or carsharing, that may not directly fall into searches returned using EV and AV keywords. EV, AV, carsharing, and ride-hailing are synergistic areas within the ground transportation field, given interest in using future AVs as an electric fleet that operates as a carsharing or ride-hailing service. However, a simple search of “ridesharing” on Scopus of publications published since 2015 conducted in September 2020 returned over 2500 publications. Thus, we opted to use a more directed approach by carefully reviewing titles and abstracts from selected journals to identify papers in the ground transportation literature that showed potential for having ideas, concepts, methods, or results that could inform UAM research.
为了确定相关的电动车和自动驾驶文章，我们回顾了 2015 年 1 月至 2020 年 6 月期间交通领域的重要期刊的目录，并根据标题和摘要确定了相关的文章。我们明确决定不使用关键词搜索来进行分析的这一部分，这样我们可以浏览标题并确定与城市空中交通研究相关的出版物，例如打车或共享汽车，这些出版物可能不会直接通过使用电动车和自动驾驶关键词返回。电动车、自动驾驶、共享汽车和打车是地面交通领域内的协同领域，因为人们有兴趣将未来的自动驾驶车辆作为电动车队运营的共享汽车或打车服务。然而，仅在 2020 年 9 月在 Scopus 上进行的自 2015 年以来发表的出版物的“打车”简单搜索就返回了超过 2500 篇文章。因此，我们选择了更有针对性的方法，通过仔细审查选定期刊的标题和摘要，确定地面交通文献中可能具有为城市空中交通研究提供思路、概念、方法或结果的论文。

Given our overarching objective in comparing the EV/AV and UAM fields is to glean insights from the EV/AV areas that may be applicable to the UAM area, we excluded some papers in the EV/AV areas that were not directly applicable to the UAM field. For example, papers that discuss strategies for safely merging AV ground vehicles into traffic are not applicable to UAM given UAM has another dimension for conflict avoidance and different traffic management rules than ground transportation modes. Similarly, when doing a detailed analysis of a particular area (such as demand segmentation), we tagged all articles that fit into the category, but then focused our in-depth discussion on the subset of articles most relevant to UAM (e.g., we exclude a discussion of how EV vehicle characteristics like acceleration influence EV purchases).
鉴于我们在比较电动汽车/自动驾驶和城市空中交通领域的总体目标是从电动汽车/自动驾驶领域中获取对城市空中交通领域有用的见解，我们排除了一些与城市空中交通领域直接不相关的电动汽车/自动驾驶领域的论文。例如，讨论如何将自动驾驶地面车辆安全地融入交通流的论文对城市空中交通来说并不适用，因为城市空中交通有另一个冲突避免维度和不同的交通管理规则。同样，在对特定领域进行详细分析时（如需求细分），我们标记了所有符合该类别的文章，但我们将重点放在与城市空中交通最相关的文章子集上进行深入讨论（例如，我们排除了关于电动汽车特性如加速度如何影响电动汽车购买的讨论）。

We reviewed articles from the following journals—the number of articles in total and those we included in our analysis are shown in parentheses: Transportation Research Part A (1392 published; 125 inventoried); Transportation Research Part B (970 published; 62 inventoried); Transportation Research Part C (1971 published; 100 inventoried); Transportation Research Part D (1281 published; 124 inventoried), Transportation (545 published; 51 inventoried); and Transportation Science (444 published; 16 inventoried).
我们回顾了以下期刊的文章 - 总共发表的文章数和我们在分析中包含的文章数分别如下：《交通研究 A 部分》（发表 1392 篇；纳入分析的 125 篇）；《交通研究 B 部分》（发表 970 篇；纳入分析的 62 篇）；《交通研究 C 部分》（发表 1971 篇；纳入分析的 100 篇）；《交通研究 D 部分》（发表 1281 篇；纳入分析的 124 篇）；《交通》（发表 545 篇；纳入分析的 51 篇）；以及《交通科学》（发表 444 篇；纳入分析的 16 篇）。

The final number of articles we identified includes 312 for UAM and 478 for EV/AV research. For each of the 790 articles, we identified research themes by associating up to six keywords based on a review of the abstracts (or where unclear, a review of the articles). For each publication, we recorded author and publication information. Information for each of these 790 articles, including DOI links, are included in an Excel sheet as a supplemental document to this paper. Co-author Garrow, an expert in travel behavior modeling from civil engineering, tagged the articles related to EVs and AVs, and co-author German, an expert in aircraft design from aerospace engineering, tagged the articles related to UAM. While the subject classifications are arguably subjective, they nonetheless enable us to identify high-level trends across the fields.
我们确定的文章总数包括 312 篇关于城市空中交通（UAM）的文章和 478 篇关于电动车辆/自动驾驶（EV/AV）研究的文章。对于这 790 篇文章中的每一篇，我们通过审查摘要（或在不清楚的情况下，审查全文）来确定研究主题，并根据最多六个关键词进行关联。对于每一篇出版物，我们记录了作者和出版信息。这 790 篇文章的信息，包括 DOI 链接，作为本文的补充文件，以 Excel 表格的形式提供。共同作者 Garrow 是土木工程中旅行行为建模方面的专家，他标记了与 EV 和 AV 相关的文章；共同作者 German 是航空航天工程中飞机设计方面的专家，他标记了与 UAM 相关的文章。虽然主题分类可能有主观性，但它们仍然使我们能够识别出各个领域的高级趋势。

3.2. UAM publications 3.2. UAM 出版物

Based on our review of UAM-related articles, we conducted a meta-analysis focused on two overarching themes: (1) categorization of the technical content of the articles, and (2) analysis of the affiliations of the authors. The former theme provides insights into the breadth and depth of the topics addressed in UAM research, and the latter provides insights into what nations, organizations, and individuals are actively focused on UAM research.
根据我们对与 UAM 相关的文章的审查，我们进行了一项以两个主题为重点的元分析：（1）对文章技术内容的分类，以及（2）对作者所属机构的分析。前一个主题提供了关于 UAM 研究所涉及的主题广度和深度的见解，而后一个主题则提供了关于哪些国家、组织和个人正在积极关注 UAM 研究的见解。

To categorize the content in the UAM-related articles, we first identified low-level topic categories that were present in multiple articles, and we created corresponding content tags. In defining these categories, we were guided in part by our knowledge of new technical topic areas related to eVTOL aircraft that are being actively addressed within the UAM community, e.g. “Distributed Electric Propulsion” and “Aero-Propulsive Interactions.” We then grouped related low-level tags hierarchically under higher-level categories associated with traditional research disciplines related to aircraft technology and operations, e.g. “Propulsion,” “Aerodynamics,” and “Simulation.” Finally, we grouped these higher-level categories into two overarching categories: “Aircraft Technology” and “Market and Operations.” The resulting categorization reflects our attempt to identify and group common themes in UAM research cogently; however, we do not claim that the categorization is mutually exclusive, collectively exhaustive, or unequivocal.
为了对与城市空中交通（UAM）相关的文章内容进行分类，我们首先确定了多篇文章中存在的低级主题类别，并创建了相应的内容标签。在定义这些类别时，我们部分地参考了我们对与电动垂直起降飞行器相关的新技术主题领域的了解，这些领域正在 UAM 社区中积极探讨，例如“分布式电动推进”和“空气推进相互作用”。然后，我们按照与飞机技术和运营相关的传统研究学科将相关的低级标签按层次分组，例如“推进系统”、“空气动力学”和“仿真”。最后，我们将这些高级类别分为两个总体类别：“飞机技术”和“市场与运营”。这种分类反映了我们试图明确地识别和分组 UAM 研究中的共同主题的努力；然而，我们并不声称这种分类是相互独立、完全穷尽或明确无误的。

The hierarchical categories are shown in Fig. 3. The numbers in parentheses indicate the number of articles with lower-level tags assigned to the corresponding category. The number of articles indicated for each higher-level parent category are summative of all children tags for the category. Note that any one article is likely to have been assigned more than one tag based on the breadth of topics covered in the article. The individual low-level content tags corresponding to the overall categories are not shown in Fig. 3 to limit the size of the figure; however, these tags are provided in the spreadsheet provided as supplemental material to this article.
图 3 显示了层次分类。括号中的数字表示分配给相应类别的较低级别标签的文章数量。每个较高级别父类别所示的文章数量是该类别的所有子标签的总和。请注意，根据文章涵盖的主题广度，任何一篇文章很可能被分配多个标签。图 3 中未显示与整体类别相对应的个别低级内容标签，以限制图的大小；然而，这些标签在提供给本文的补充材料电子表格中提供。

1. Download : Download high-res image (166KB)
   下载：下载高分辨率图像（166KB）
2. Download : Download full-size image
   下载：下载完整尺寸的图片

Fig. 3. UAM themes from AIAA and Scopus search, January 2015 to June 2020.
图 3. 从 AIAA 和 Scopus 搜索中得出的 UAM 主题，时间范围为 2015 年 1 月至 2020 年 6 月。

The first observation from this analysis is that current articles on UAM have a nearly even split of content related to “Aircraft Technology” (295 papers) and “Market and Operations” (248 papers). This thematic balance likely reflects an understanding within the community of the “chicken-and-egg” issue associated with the emergence of UAM, i.e., aircraft must be technically capable of serving the missions required for profitable large-scale UAM operations, and a market must exist for the types of missions and operations that can be supported given the technological limitations of emerging aircraft. A concrete example of this interplay is related to eVTOL aircraft with battery electric propulsion. These aircraft have the capability of being much quieter and more economical than current generation helicopters, potentially allowing widespread operations in urban environments at low ticket prices. However, battery electric eVTOL aircraft have very limited range and speed capability because of the low specific energy of current and near-term batteries, potentially limiting the potential for the aircraft to serve an adequate network of origins and destinations and to offer adequate travel time savings compared to other modes when trip times are dominated by ingress and egress on short-ranged flights.
这项分析的第一个观察结果是，目前关于城市空中交通（UAM）的文章几乎平均分布在“飞行器技术”（295 篇论文）和“市场与运营”（248 篇论文）两个主题上。这种主题平衡可能反映了社区对于 UAM 出现的“先有鸡还是先有蛋”的问题的理解，即飞行器必须在技术上具备为盈利的大规模 UAM 运营提供所需任务的能力，同时市场必须存在于能够在新兴飞行器技术限制下支持的任务和运营类型。一个具体的例子是与电池电动垂直起降飞行器（eVTOL）相关的相互作用。这些飞行器具有比当前一代直升机更安静和更经济的能力，潜在地可以以低票价在城市环境中进行广泛运营。 然而，由于当前和近期电池的比能量较低，电池电动垂直起降飞行器的航程和速度能力非常有限，这可能限制了飞行器在为足够的起点和终点网络提供服务以及在行程时间主要由短程飞行的进出口所主导时，相比其他交通方式提供足够的旅行时间节约的潜力。

Within the “Aircraft Technology” category, the majority of papers had content related to “Propulsion” (82 papers) and “Aircraft Design and Performance” (92 papers). The “Propulsion” category includes papers with content related to new propulsion architectures relevant for UAM, including “Electric Propulsion” (battery powered propulsion; 29 papers), “Distributed Electric Propulsion” (multiple electric motors powering rotors or fans located in multiple locations on the aircraft; 9 papers), and “Hybrid Propulsion” (propulsion powered by both batteries and a combustion engine; 18 papers) as well as papers with a focus on associated propulsion component technologies such as “Battery” (12 papers) and “Fuel Cell” (4 papers). The majority of the papers in the “Aircraft Design and Performance” category is related to “Design Methods” (30 papers) or present a “Concept Study” of the design and performance of a novel UAM aircraft configuration (41 papers). The “Aircraft Technology” category has a significant number of papers related to “Aerodynamics” (43 papers), as well, with many papers focused specifically on the aerodynamics associated with eVTOL aircraft and other distributed propulsion configurations, i.e. “Propeller, Rotor, Ducted Fan” (16 papers) and “Aero-Propulsive Interaction” (7 papers).
在“飞机技术”类别中，大部分论文内容与“推进系统”（82 篇论文）和“飞机设计与性能”（92 篇论文）有关。 “推进系统”类别包括与城市空中交通相关的新型推进系统架构的论文，包括“电动推进”（电池驱动推进系统；29 篇论文），“分布式电动推进”（多个电动马达驱动位于飞机多个位置的旋翼或风扇；9 篇论文）和“混合推进”（电池和燃烧发动机共同驱动推进系统；18 篇论文），以及关注相关推进系统组件技术的论文，如“电池”（12 篇论文）和“燃料电池”（4 篇论文）。 “飞机设计与性能”类别中大部分论文与“设计方法”（30 篇论文）有关，或者是对新型城市空中交通飞机配置的设计和性能进行“概念研究”（41 篇论文）。 “飞行器技术”类别中有大量与“空气动力学”相关的论文（43 篇），同时也有许多论文专注于与电动垂直起降飞行器和其他分布式推进配置相关的空气动力学，即“螺旋桨、旋翼、风扇罩”（16 篇）和“空气动力相互作用”（7 篇）。

The large number of papers focused on novel propulsion—especially electric propulsion—and ways to design novel aircraft with new propulsion technology is not surprising. Indeed, electric propulsion is widely recognized as one of the underpinning technical enablers of aircraft capable of serving the UAM market. What is more surprising is that other key technical disciplines for enabling UAM aircraft such as “Autonomy” (11 papers), “Acoustics” (16 papers), and “Safety” (18 papers) have relatively few papers and are arguably underrepresented relative to their importance to the field. The relatively few papers focused on autonomy is likely a result of the very broad character of autonomy research, which focuses on technical fundamentals, as well as a myriad of application domains, including ground AVs. This breadth has likely resulted in few researchers focusing on autonomy research specifically for the emerging field of UAM. Additionally, aviation is a highly-regulated industry with inherent skepticism about the potential of autonomy for replacing pilots in the near future; this viewpoint has led to research in simplified vehicle operations (SVO) focused on enabling piloted aircraft with increasingly automated but not autonomous systems for flight control and navigation (Goodrich and Moore, 2015). The relatively few papers in the “Acoustics” category may result from the ramp-up of the research community to develop fundamentally new foundational computational tools and appropriate metrics for UAM aircraft noise, which differ substantively from traditional aviation noise metrics (Josephson, 2017). The few papers in the “Safety” category may be a result of the need to make initial research progress to address the novelty of UAM aircraft, which require envisioning entirely new paradigms for achieving safety. For example, the simple rotor systems in eVTOL aircraft do not typically offer the potential for “autorotation” for safe descents after an engine failure that is available to helicopters; instead eVTOL aircraft are designed with multiply-redundant powertrain components to prevent a complete propulsion failure in flight (Fredericks, 2016). The few papers in these categories may represent an opportunity for researchers to have impact by engaging in UAM research in these critically important research fields.
大量的论文关注新型推进技术，尤其是电动推进技术，以及设计新型飞机的方法，这并不令人意外。事实上，电动推进被广泛认为是能够为城市空中交通市场提供服务的飞机的基础技术之一。更令人惊讶的是，其他关键的技术学科，如“自主性”（11 篇论文）、“声学”（16 篇论文）和“安全性”（18 篇论文），相对较少，可以说在该领域的重要性相对低估。相对较少关注自主性的论文可能是自主性研究非常广泛的结果，它关注技术基础以及包括地面自动驾驶车辆在内的众多应用领域。这种广度可能导致很少的研究人员专门关注自主性研究，特别是针对新兴的城市空中交通领域。 此外，航空业是一个受到高度监管的行业，对于自主飞行取代飞行员的潜力存在固有的怀疑；这种观点导致了简化飞行器操作（SVO）的研究，重点是为了使有驾驶员的飞行器具备越来越自动化但不是完全自主的飞行控制和导航系统（Goodrich 和 Moore，2015）。在“声学”类别中相对较少的论文可能是由于研究界正在加紧开发基础性的全新计算工具和适用于城市空中交通（UAM）飞行器噪音的合适指标，这些指标与传统航空噪音指标有实质性的区别（Josephson，2017）。在“安全”类别中的少数论文可能是因为需要取得初步的研究进展来应对 UAM 飞行器的新颖性，这需要构想完全新的实现安全的范式。 例如，eVTOL 飞行器中的简单转子系统通常不具备直升机所具备的发动机故障后安全下降的“自转”能力；相反，eVTOL 飞行器采用了多重冗余的动力传动组件，以防止飞行中完全失去推进力（Fredericks，2016）。这些类别中的少数论文可能代表了研究人员在这些至关重要的研究领域参与城市空中交通研究的机会。

Within the “Market and Operations” category, the majority of papers had content related to “Air Traffic Management” (83 papers) and “Aviation Operations” (80 papers). The “Air Traffic Management” category includes papers focused on topics such as exploring paradigms for integrating large volumes of UAM air traffic within the existing NAS (Mueller et al., 2017, Thipphavong et al., 2018), constraints on UAM operations based on current operations at major airports (Vascik and Hansman, 2017, Vascik et al., 2018), and assessing and increasing airspace density and throughput for UAM operations (Goodrich and Barmore, 2018, Lowry, 2018). The “Aviation Operations” category includes papers with a focus on topics of economic and practical interest to UAM air carriers, including flight planning (Stouffer and Kostiuk, 2020) flight scheduling and dispatch (Roy et al., 2020, Shihab et al., 2019, Shihab et al., 2020), concepts of operations (Nneji et al., 2017; Kotwicz et al., 2019), and issues associated with electric aircraft recharging for flight operations (Hamilton and German, 2019, Shihab et al., 2020).
在“市场与运营”类别中，大部分论文内容与“空中交通管理”（83 篇论文）和“航空运营”（80 篇论文）相关。 “空中交通管理”类别包括关注将大量城市空中交通整合到现有国家空中交通系统中的范式探索（Mueller 等，2017 年；Thipphavong 等，2018 年），基于主要机场当前运营情况对城市空中交通运营的限制（Vascik 和 Hansman，2017 年；Vascik 等，2018 年），以及评估和提高城市空中交通运营的空域密度和吞吐量（Goodrich 和 Barmore，2018 年；Lowry，2018 年）。 “航空运营”类别包括关注对城市空中运营商具有经济和实际意义的主题的论文，包括飞行计划（Stouffer 和 Kostiuk，2020 年）、飞行调度和派遣（Roy 等，2020 年；Shihab 等，2019 年；Shihab 等，2020 年）、运营概念（Nneji 等，2017 年；Kotwicz 等，2019 年）以及与电动飞机充电飞行运营相关的问题（Hamilton 和 German，2019 年；Shihab 等，2020 年）。

A significant number of papers in the “Market and Operations” category was focused on transportation studies and research to assess the potential of UAM for providing an effective and scalable means of reducing travel time in cities, assessments which lend to understanding the market potential of UAM and its value to society. Papers in the “Air Transportation Studies” category (39 papers) assessed specific types of novel UAM aircraft in on-demand or scheduled service, typically through the lens of one or several operational case studies in example cities, and papers in the “Multimodal Transportation Studies” category (14 papers) assessed connections between UAM and other transportation modes such as cars or public transport or at least discussed differences between UAM and other transportation modes. Examples of applied transportation studies include a series of papers on “suburban air mobility” with electric short takeoff and landing (eSTOL) in the south Florida region (Wei et al., 2018, Robinson et al., 2018, Justin and Mavris, 2019, Somers et al., 2019). Papers in the “Demand” category (20 papers) focused on topics such as assessing the potential market size for UAM and other forms of on-demand air mobility based on census data and choice models (Kreimeier et al., 2018, Roy et al., 2020, Ploetner et al., 2020), stated preference surveys to assess UAM demand (Binder et al., 2018, Garrow et al., 2020, Fu et al., 2019), and agent-based demand simulation (Rothfeld et al., 2018, Fu et al., 2019, Ploetner et al., 2020). Finally, the “Infrastructure” category (12 papers) includes papers focused on optimization-based site selection of new vertiports (Daskilewicz et al., 2018) and STOLports (Wei et al., 2020) to serve the maximum demand, as well as papers that assess capacity constraints of vertiports (Vascik and Hansman, 2019, Maheshwari et al., 2020).
“市场与运营”类别中的大量论文集中在交通研究上，评估无人机空中交通（UAM）在提供有效和可扩展的减少城市出行时间方面的潜力，这些评估有助于了解 UAM 的市场潜力和对社会的价值。在“航空运输研究”类别（39 篇论文）中，通过在示范城市中进行一个或多个运营案例研究的视角，评估了特定类型的新型 UAM 飞行器在按需或定期服务中的表现，而“多模式交通研究”类别（14 篇论文）则评估了 UAM 与其他交通方式（如汽车或公共交通）之间的联系，或至少讨论了 UAM 与其他交通方式之间的差异。应用交通研究的例子包括一系列关于在佛罗里达南部地区使用电动短距起降（eSTOL）的“郊区空中移动”论文（Wei 等，2018 年；Robinson 等，2018 年；Justin 和 Mavris，2019 年；Somers 等，2019 年）。 “需求”类别的论文（20 篇）关注的主题包括基于人口普查数据和选择模型评估 UAM 和其他形式的按需空中移动的潜在市场规模（Kreimeier 等，2018 年，Roy 等，2020 年，Ploetner 等，2020 年），声明偏好调查评估 UAM 需求（Binder 等，2018 年，Garrow 等，2020 年，Fu 等，2019 年），以及基于代理人的需求模拟（Rothfeld 等，2018 年，Fu 等，2019 年，Ploetner 等，2020 年）。最后，“基础设施”类别的论文（12 篇）包括关注新垂直港口（Daskilewicz 等，2018 年）和 STOL 港口（Wei 等，2020 年）的基于优化的选址，以满足最大需求的论文，以及评估垂直港口容量限制的论文（Vascik 和 Hansman，2019 年，Maheshwari 等，2020 年）。

In our meta-analysis of author affiliations, the 251 UAM-related articles from AIAA consisted of a total of 862 listed authors, many of whom were listed on multiple papers, resulting in 554 unique authors. Among the 554 unique authors, 44 percent are affiliated with an academic institution, and 31 percent are associated with NASA. The remaining 25 percent of authors are associated with U.S.-based and international companies and research agencies. The majority of authors in the AIAA database (83 percent) are affiliated with institutions in the U.S., and the country with the second-highest representation (7 percent) is Germany. As these statistics reveal, the majority of UAM research has been conducted by the U.S., and NASA has played a critical role in this research.
在我们对作者机构的元分析中，AIAA 的 251 篇与 UAM 相关的文章共涉及 862 位列出的作者，其中许多人在多篇文章中都有名字，因此总共有 554 位独特的作者。在这 554 位独特的作者中，有 44%与学术机构有关联，31%与 NASA 有关联。剩下的 25%的作者与美国和国际公司以及研究机构有关联。AIAA 数据库中的大多数作者（83%）与美国的机构有关联，其次是德国（7%）。正如这些统计数据显示的那样，大部分 UAM 研究是由美国进行的，NASA 在这项研究中起到了关键作用。

A similar meta-analysis was conducted with UAM articles returned from the Scopus search with AIAA publications excluded. The 61 UAM-related articles from the Scopus search consisted of a total of 175 listed authors and 141 unique authors. Of all 141 unique authors, 66 percent are affiliated with an academic institution, and 16 percent are affiliated with NASA. The remaining 18 percent of authors are affiliated with U.S.-based and international companies and research agencies. Similar to the results seen in the AIAA database, the majority of authors (52 percent) are affiliated with institutions in the U.S., and the country with the second-most representation in the Scopus search is Germany (20 percent). The country with the third-most representation is the Republic of Korea (4 percent). These statistics confirm the trends seen in the AIAA search—UAM research has been concentrated primarily among U.S.- and German-based researchers, and NASA has played a critical role.
从 Scopus 搜索中排除 AIAA 出版物，对返回的 UAM 文章进行了类似的元分析。来自 Scopus 搜索的 61 篇与 UAM 相关的文章共有 175 位列出的作者和 141 位独立作者。在这 141 位独立作者中，66%隶属于学术机构，16%隶属于 NASA。其余 18%的作者隶属于美国和国际公司以及研究机构。与 AIAA 数据库中的结果类似，大多数作者（52%）隶属于美国的机构，Scopus 搜索中代表性第二高的国家是德国（20%）。代表性第三高的国家是韩国（4%）。这些统计数据证实了在 AIAA 搜索中观察到的趋势——UAM 研究主要集中在美国和德国的研究人员之间，并且 NASA 发挥了关键作用。

3.3. Ground transportation publications
地面交通出版物

To categorize the content of ground-transportation articles, we first identified broad topics. Many of these topics are overlapping and represent envisioned synergies across new technologies. For example, papers that discuss a future in which a fleet of AVs operate on batteries would be classified under the high-level categories of “Electric Vehicles” and “Autonomous Vehicles.” Once we identified broad topics, we tagged themes within each topic area that were potentially relevant for UAM research. The content tags are shown in Fig. 4. Later sections present our review of these lower-level tags in depth, so we restrict our discussion here to one key observation: within the top-tier transportation journals identified on Hensher’s list (2019), there were only four articles published on UAM. This highlights the opportunity for the transportation planning community to take a more active role in research related to the design and operations of UAM systems and apply insights they have gained through related research in the EV and AV fields to UAM.
为了对地面交通文章的内容进行分类，我们首先确定了广泛的主题。其中许多主题是重叠的，并代表了新技术之间的协同效应。例如，讨论未来一支由电池驱动的自动驾驶车队的论文将被归类为“电动车辆”和“自动驾驶车辆”这两个高级分类。一旦我们确定了广泛的主题，我们在每个主题领域内标记了可能与 UAM 研究相关的主题。内容标签如图 4 所示。后面的章节将详细介绍我们对这些低级标签的审查，因此我们在这里限制我们的讨论只针对一个关键观察：在 Hensher（2019）的列表中确定的顶级交通期刊中，只有四篇关于 UAM 的文章发表。这凸显了交通规划社区在 UAM 系统的设计和运营方面扮演更积极角色的机会，并将他们在电动车辆和自动驾驶领域的相关研究所获得的见解应用于 UAM。

1. Download : Download high-res image (214KB)
   下载：下载高分辨率图像（214KB）
2. Download : Download full-size image
   下载：下载完整尺寸的图片

Fig. 4. Themes from transportation journals’ table of contents search, January 2015 to June 2020.
图 4. 从交通期刊的目录搜索中得出的主题，时间范围为 2015 年 1 月至 2020 年 6 月。

The 478 ground transportation articles from the journals Transportation Research Part A (TR–A), TR–B, TR–C, TR–D, Transportation Science, and Transportation consisted of a total of 1,594 listed authors, many of whom were listed on multiple papers, resulting in 1,154 unique authors. Among the 1,154 unique authors, 84 percent are affiliated with an academic institution. The remaining 16 percent of authors are associated with U.S.-based and international companies and research agencies. Among authors associated with academic institutions, 26 percent are affiliated with institutions in the U.S. The country with the second-highest representation (13 percent) in the ground transportation journal database is China, closely followed by Germany (9 percent). As these statistics reveal, the majority of ground transportation research has been conducted by the U.S., but the authors are much more diverse in their affiliated countries than the UAM authors. Ground transportation authors are also much more commonly affiliated with academic institutions compared to UAM authors.
来自《交通研究 A 部分》（TR-A）、《交通研究 B 部分》（TR-B）、《交通研究 C 部分》（TR-C）、《交通研究 D 部分》（TR-D）、《交通科学》和《交通》期刊的 478 篇地面交通文章共涉及 1594 位作者，其中许多作者在多篇文章中都有署名，因此共有 1154 位独特的作者。在这 1154 位独特的作者中，84%与学术机构有关联。剩下的 16%的作者与美国和国际公司以及研究机构有关联。在与学术机构有关联的作者中，有 26%与美国的机构有关联。在地面交通期刊数据库中，第二大代表国家（占 13%）是中国，紧随其后的是德国（占 9%）。正如这些统计数据所显示的，大部分地面交通研究是由美国进行的，但作者的所属国家比 UAM 作者更多样化。与 UAM 作者相比，地面交通作者更常常与学术机构有关联。

4.1. Review of UAM demand studies
4.1. UAM 需求研究的回顾

This section reviews three types of demand studies that have been conducted by UAM researchers: global market studies that have ranked cities worldwide for their potential to offer UAM services, studies that have compared potential travel times savings with UAM against other modes, and survey-based research.
本节回顾了由 UAM 研究人员进行的三种需求研究类型：全球市场研究，对全球城市进行排名以评估其提供 UAM 服务的潜力；将 UAM 与其他交通方式进行比较，评估潜在的出行时间节约；以及基于调查的研究。

4.2. Review of EV and AV demand studies
4.2. 电动车和自动驾驶需求研究的回顾

Within the ground transportation literature, there have been more than 200 studies over the past five years that have focused on demand. Some of these studies focus on understanding how demand for EV, AV, carsharing, and/or ride-hailing services varies as a function of sociodemographic and socioeconomic (SED) characteristics, as well as different attitudes, beliefs, and personality factors. About 50 studies within the ground transportation literature have focused on how adoption of new EV and AV technologies will increase over time. These studies include an assessment of barriers to adoption, analysis of the differences between early adopters and late adopters, and extension and application of different theoretical frameworks used to predict the timing of adoption across a population. Finally, there are two topics that have been explored in the ground transportation literature that are particularly relevant for UAM: studies that have examined individuals’ value of time, defined as the amount of money individuals are willing to spend for travel time savings, and studies that have examined individuals’ willingness to ride in vehicles with individuals they know or strangers. This section reviews these demand-related topics in depth.
在地面交通文献中，过去五年已经有 200 多项研究集中在需求方面。其中一些研究着重于理解电动汽车（EV）、自动驾驶汽车（AV）、共享汽车和/或打车服务的需求如何随社会人口特征、经济社会特征以及不同的态度、信念和个性因素而变化。地面交通文献中约有 50 项研究集中在新的电动汽车和自动驾驶技术的采用将如何随时间增加。这些研究包括评估采用障碍、分析早期采用者和晚期采用者之间的差异，以及扩展和应用用于预测整个人群采用时间的不同理论框架。最后，地面交通文献中探讨的两个与城市空中交通（UAM）特别相关的主题是：研究个人时间价值，即个人愿意为节省旅行时间而花费的金额；以及研究个人愿意与认识的人或陌生人一起乘坐车辆的意愿。 本节深入讨论了这些与需求相关的主题。

4.3. Bringing it all together—Demand modeling insights and research directions for UAM

As seen from the literature review, demand modeling within the UAM and EV/AV domains have focused on different objectives. Within the UAM field, the primary focus has been on determining if UAM is a viable concept—e.g., will enough people be willing to fly in these new air taxis and can the service be supported across different cities? Within the ground transportation field, EVs, AVs with lower levels of automation, and sharing services have already been implemented, allowing researchers to focus on understanding SED characteristics of early adopters or how individuals respond to different policy incentives and operational policies.

To the extent that individuals who are interested in EV, AV, and/or sharing modes will also be interested in air taxis, we would expect that early adopters of air taxis will be more likely to be male, have higher incomes, have pro-environmental attitudes and/or be tech-savvy, technology-oriented lifestyles and be the first to try out new products. These expectations have been confirmed in surveys of U.S. commuters by Boddupalli et al. (2020) and Garrow, Roy, and Newman (2020).

The EV and AV literature have several findings that are relevant for the UAM community. To date, there has been a significant amount of research in the EV and AV literature that has looked at the timing of when adoption occurs, but only one paper in the UAM area, by Al Haddad et al. (2020). These technology adoption models can provide valuable information on the role of trust, safety, and perceived usefulness on the adoption of UAM. The literature across both the air and ground transportation areas show mixed reactions in the population with respect to autonomy. Finding ways to increase individuals’ trust in autonomy would be a valuable direction for future research. For example, we may find that it is important to provide demonstrations of what it would be like to fly in a UAM using virtual reality and/or to provide safety information to increase individuals’ comfort levels with the new technology. The role of social effects (like trusting perceptions of friends and family) has been shown to play a role in adoption of ground vehicle technologies and could be investigated in the context of UAM.

Unlike with ground transportation modes, individuals are more likely to expect to travel with strangers in an aircraft. It is, thus, unclear whether the same effects seen for ride-hailing services will apply to UAM. One study, by Garrow, Roy, and Newman (2020) did find that younger commuters were less likely to take a UAM with strangers compared to older commuters. However, there is a research need to understand if the willingness to travel with strangers in a UAM aircraft varies across nations and trip purposes.

Perhaps one of the most interesting findings from the AV literature is that the VOT for commuters will decrease when ground AVs enter the market due to the ability for commuters to use their time more productively. From the UAM perspective, this is important as it suggests that AVs will compete more heavily with air taxis than with conventional autos and that additional travel time savings will be required for the air taxi mode relative to the AV. Potential productivity gains in an AV compared to an air taxi have not been explored in the literature, and there is a need to determine what levels of productivity would be achievable in a UAM vehicle and how productivity varies as a function of ride quality, trip duration, and other factors. Given the VOT decreases seen for AV ground research, better understanding of VOT decreases for UAM vehicles—particularly as they relate to commute trips—is an important area of future research.

Another interesting avenue for future research would be to explore how AVs and air taxis will compete across different trip purposes as an air taxi system evolves and adoption rates increase across both new modes.

5.1. Review of UAM infrastructure studies

Several UAM researchers have focused on infrastructure-related issues, mostly in the context of UAM operations. This section highlights the tight couplings researchers have observed among vertiport placement, operations, demand, and energy requirements.

5.2. Review of EV and AV infrastructure studies

This section provides an overview of the types of questions that researchers and policy makers have investigated when integrating a new mode into the existing ground transportation network, and the role of parking availability on mode choice. This discussion is followed by a detailed review of a topic that is particularly relevant to UAM: integration with the electric grid.

5.3. Bringing it all together—Infrastructure insights and research directions for UAM

Across both the UAM and EV/AV literature, station placement has been shown to be a critical factor influencing overall system performance. The sheer volume of publications in the EV area that have focused on charging infrastructure or charging type is noteworthy—about one out of every five EV papers we inventoried addressed these topics, as indicated in Fig. 3. Within the EV community, significant attention has been placed on understanding individuals’ charging behavior and strategies for shifting charging patterns to reduce the peak period load on the electric grid. This is relevant from a UAM perspective, as it suggests that the transportation community is already experiencing challenges associated with charging a ground EV fleet. Some EV research has suggested that current electric grids won’t be able to support future EV ground vehicle charging needs. Needless to say, if we are not in a position to handle charging of a ground EV fleet, how are we going to handle charging a UAM fleet that will likely require even faster charge times? There is a clear research need to better understand the power profiles of UAM fleets and develop strategies for how to optimally charge UAM fleets without overwhelming the electric grid. Another interesting topic would be to jointly examine the power profiles for UAM fleets and EV ground fleets, as both technologies will be competing for a limited amount of electricity.

As will become more evident in the next section that focuses on operations, the placement of vertiports is closely coupled with operations and other factors. All of this points to the need to develop high-fidelity simulation models for UAM operations that capture interactions among vertiport locations, vertiport topology, demand, pricing, dispatching, and airspace restrictions.

In comparing the UAM and AV/EV literatures, we could find no mention of the role of parking availability on air taxi mode choice, and we suggest this could be an interesting factor to include in future air taxi mode choice studies, particularly studies that included ride-hailing, air taxis, and traditional autos as potential modes. As we integrate UAM into our cities, it will be important to work with local planning organizations to ensure that any infrastructure investments that require public funding align with regulations these organizations need to follow.

6.1. Review of UAM operations studies

As the UAM community designs an air taxi system capable of high-volume throughput integrated in urban areas, many operations-related questions arise. One of the first steps in the analysis process is to design a concept of operations (ConOps), which is essentially a plan for how UAM operations can be safely integrated into the national airspace system. In June 2020, NASA released its ConOps vision, which includes UAM corridors in the sky in which aircraft could operate without the direct involvement of air traffic control (ATC) (Bradford, 2020). Given the importance of ensuring safe operations within the existing NAS, it is not surprising that the UAM community has focused significant attention on ATC-related issues (as shown in the meta-analysis, Fig. 2).

Given a concept of operations, researchers can assess whether a particular aircraft design can successfully and economically perform a given mission, and if it cannot, make modifications to the aircraft design (e.g., see Clarke et al. 2019). To determine whether a mission can be performed successfully for an electric-powered aircraft, researchers need to model the mission’s power and energy requirements, which imply the peak current and total capacity required by the battery. For example, Kulkarni et al. (2018) developed an on-board battery monitoring and prognostic architecture for batteries on electric-propulsion aircraft. Alnaqeb et al., 2018, Prabhakar et al., 2020 developed models to predict mission-based energy and performance metrics; Donateo and Ficarella (2020) proposed a modeling approach for the degradation of the battery performance during its aging; and Pradeep and Wei (2019) developed energy-efficient trajectory plans for a multirotor eVTOL. Hamilton and German, 2017, Hamilton and German, 2019 optimized airspeeds for electric aircraft operations to maximize energy feasibility in the schedule by balancing energy expended during cruise and energy replenished during recharge.

Several researchers have investigated the relationships among aircraft design parameters and operational requirements such as cruise speed and hover time. For example, using the Uber eCRM-001 eVTOL common reference model (Uber Elevate, 2020a), Ha et al. (2020) jointly optimized aircraft design parameters in addition to operational parameters to achieve a 9.66 percent decrease in required hover power. Other researchers have examined the potential for retrofitting existing aircraft with an electric propulsion system to determine if such aircraft could profitably operate for pilot-training applications (Olson, 2015) or short-haul UAM intracity commuter trips in U.S. cities (Kotwicz et al., 2019).

UAM clearly will not be successful without a ConOps that safely integrates aircraft into the NAS and aircraft that can complete the required missions. Thus, it is not surprising that much of the research by the UAM community has been focused on mission performance and related areas, such as battery design and battery modeling. However, as the vision for UAM ConOps and aircraft designs has become clearer, the UAM community is starting to focus on more complex operational issues that include dispatching algorithms and pricing approaches, which are similar topics explored by the EV and AV communities. For example, Roy et al. (2019) examined how existing infrastructure, resources, and operational strategies could be leveraged with improvements in battery and autonomy for regional air mobility. Roy, Crossley, et al. (2018) jointly optimized aircraft designs, operations, and revenue management, and Roy et al. (2020) developed a dispatch model to optimally schedule UAM flights for a shuttle service to an airport that has both scheduled and on-demand customers. Shihab et al. (2019) developed a model to decide whether to offer on-demand or scheduled flights, and how to dispatch the fleet and schedule operations based on simulated market demand. Munari and Alvarez (2019) assigned aircraft to on-demand requests while accounting for maintenance events, allowing flight upgrades in order to reduce operational and repositioning costs. Narkus-Kramer et al. (2016) examined trade-offs associated with battery-powered, remotely piloted semi-autonomous personal aircraft, and found that profitability is closely tied to high network utilizations (which result in fewer deadhead and repositioning flights) and high daily utilization (or higher average hours flown). Finally, Stouffer and Kostiuk (2020) designed a dispatching tool for UAM operations that enables a dispatcher to plan a UAM flight and check for issues before filing a flight plan.

The majority of the papers to date focused on dispatching and scheduling algorithms have presumed a deterministic framework, but two important consideration in UAM applications are that scheduling and dispatching algorithms may need to be done in real time or using a rolling horizon framework to account for delays and uncertainties, and these algorithms may need to be applied at a network level. As Thipphavong et al. (2018) noted, “due to limited energy reserves, UAM aircraft must have assurances prior to takeoff that their destination landing site will be available when they arrive. The tight coupling between arrivals and departures across the vertiports in a UAM network points to the possible need for continuous network-wide scheduling as a first-order control method for real-time, on-demand resource management.”

6.2. Review of EV and AV operations studies

EVs have been integrated into many communities throughout the world. As a result, researchers have been able to both develop and validate models using case studies. In the process, researchers have gained many insights regarding how system performance and profitability are affected by fleet size, demand, pricing, and reservation and dispatching strategies. Many of these insights are relevant to the UAM community, particularly given similarities in the directional demand patterns that both ground and UAM communities seek to serve.

From an operations perspective, developing strategies to serve one-way demand while maintaining profitability has been a particularly vexing problem for vehicle-sharing companies. Many travel patterns exhibit strong uni-directional flows, especially during peak periods. For example, in many cities morning rush hour traffic is created by commuters traveling from the suburbs into the city centers to work. Before COVID-19, airports that served predominately business travelers would see peaks of passengers traveling to the airport for Monday morning flights and peaks of passengers leaving the airport Thursday evening and/or Friday to return home. One-way demand patterns result in the need to increase the number of vehicles available to serve peak directional demand (e.g., see Hörl et al., 2019) and/or increase the need to reposition empty vehicles. Staging vehicles to serve peak demand and/or attempting to temporally or spatially shift demand to nearby pick-up and drop-off locations are some strategies that have been explored to serve one-way demand profitably (e.g., see Ströhle et al., 2019).

Within the ground transportation literature, many researchers have focused on the vehicle relocation problem, often in the context of one-way demand systems. Illgen and Höck (2018) provide a review of methods used to relocate vehicles in carsharing networks. Representative studies include those by Wang, Liu, and Ma (2019) and Wang, Yang, and Zhu (2018), who examined one-way electric carsharing systems; Warrington and Ruchti (2019), who studied Philadelphia’s public bike-sharing program; and Vasconcelos et al. (2017), who studied a carsharing service in Lisbon, Portugal, and found that relocating vehicles generated an additional 19–24 percent in profits for operators.

Given that “the cost associated with vehicle relocation operations represents a significant proportion of the total operating cost” (Boyacı and Zografos, 2019), many researchers have developed methods for better predicting demand and for tailoring operational strategies to minimize relocation costs while maintaining high service levels. Wen et al. (2019) examined dispatching polices with different types of demand information for an AV shared system and found that individual demand information from in-advance requests improves performance, but the degree of performance depends on the spatial disparity of requests. Boyacı and Zografos (2019) examined temporal and spatial flexibility regarding pick-up and drop-off of vehicles in a one-way electric carsharing system and found that spatial flexibility has a stronger effect than temporal flexibility, but both temporal and spatial flexibility can increase profitably of the system by serving more customers with fewer relocation needs. Hyland and Mahmassani (2018) compared different dispatching policies for an AV service and found that the optimal dispatching policy is a function of demand, with more sophisticated dispatching polices generating higher revenues during the peak demand period and simple dispatching policies (i.e., assigning passengers sequentially to nearest idle AV) working well in low demand periods. This result is consistent with the findings based on a case study of Zurich, Switzerland, that investigated different operational policies for an AV shared mobility system and found that operational policies had a significant impact on vehicle assignment and repositioning, heavily influencing system performance of wait times and cost (Hörl et al., 2019). Both Hyland and Mahmassani (2018) and Hörl et al. (2019) found that the utilization of intelligent demand forecasts and dispatching and rebalancing algorithms were crucial elements of profitability. In addition, Hyland and Mahmassani (2020) found that increases in the mean curbside pick-up time for a shared AV system significantly degrades operational performance in terms of user in-vehicle travel time and user wait time.

The role of advance reservations in the profitability of sharing services is nuanced. On one hand, advance reservations provide more certainty with respect to future demand and allow the operator to position vehicles in advance to the locations where customers have requested service. However, if operators take vehicles out of service too far in advance to guarantee availability for reservations, then vehicle utilization and the ability to serve on-demand requests may decrease, resulting in a less profitable system. As Molnar and Correia (2019) pointed out “while it is convenient for customers to be able to do one-way trips and drop off vehicles anywhere in a service area, this makes it difficult to offer reservations in advance” and there is a need to explore ways to increase advance reservation times by relocating vehicles to shortly before reservation pick-up times.

Several researchers have explicitly focused on the issue of advance reservations and traveler flexibility. Wu et al. (2019) examined the role of guaranteed advance reservations for a free-floating carsharing service in London and found that individuals are willing to pay £0.54 per journey ($0.75 USD)¹⁶ for a guaranteed advance reservation. Duan et al. (2020) examined a system in which individuals can either request immediate rides or reserve an AV taxi service in advance, and optimized a model that considers vehicle-to-passenger assignment with empty vehicle rebalancing. They found that when the number of vehicles is adequate and reservations are made further ahead of time, the completion rate of requests and revenue improve. Allahviranloo and Chow (2019) examined a system in which individuals can buy future time slots for AV and are guaranteed service. They found the spatial temporal distribution of demand impacts the solution to the fleet sizing problems.

Several researchers have jointly optimized fleet size and trip pricing for sharing systems. Xu, Meng, and Liu (2018) jointly optimized EV fleet size and trip pricing for a one-way carsharing service that considers vehicle relocation and personnel assignment based on a case study of Singapore. Jorge et al. (2015) used a theoretical case study network of 75 carsharing stations in Lisbon, Portugal, and found that trip pricing can increase profits through more balanced systems; optimal profits are on average 23 percent higher than base prices and serve 18 percent less demand.

Finally, several researchers have noted that the optimal operational policies and/or deployment of charging stations will evolve over time as demand increases. Ghamami, Zockaie, and Nie (2016) found that ignoring delay induced by charging congestion led to suboptimal configuration of charging infrastructure, with effects potentially more prominent as demand increased over time for PEVs. Wu and Sioshansi (2017) found challenges in planning placement of public fast-charging stations for EV due to uncertainty in future demand with initial expansion concentrated around the urban core. Dong, Ma, et al. (2019) found that as additional charging stations are built, the optimal locations start in central London and gradually expand out to suburban areas of London. Zhang, Schmöcker, et al. (2020) found when expanding one-way carsharing stations, demand growth is higher around transit hubs and public facilities than in residential areas.

6.3. Bringing it all Together—Operations insights and research directions for UAM

Based on prior research from the ground transportation literature, it is clear that system performance and profitability is driven by multiple factors, including the spatial and temporal distribution of demand, fleet size, pricing, and operational policies, and that there are strong couplings across these factors. As Repoux et al. (2019) eloquently stated, “The interaction between all parameters and settings in carsharing is complex and highly non-linear. It re-emphasizes the importance for any practitioner to identify the most effective elements (namely fleet size, station capacities, rental rules) as well as the ones specific to the system’s environment and demand.” From a UAM perspective, this highlights a critical need to jointly optimize interactions among fleet, demand, pricing, and dispatching policies. The fact that many use cases for UAM (such as commuting and trips to the airport) exhibit strong directional or one-way demand patterns will likely put further pressure on the profitability of UAM networks. One key difference between the UAM and EV/AV communities relates to the need for real-time optimization and dispatching algorithms. The penalty for running out of battery energy is much more severe in air applications than ground applications; simply stated, an eVTOL aircraft cannot run out of battery power for safety reasons. Consequently, approaches that synchronize takeoffs and landings at a vertiport in real time or under a rolling horizon framework will likely be much more critical (e.g., see Kleinbekman et al., 2020).

The experiences from the ground transportation literature with respect to the potential reduction in utilization caused by guaranteeing advance reservations is particularly relevant for the UAM community, given many customers may expect a high level of availability for their flights. Results from the ground transportation literature that find profitability can be significantly increased by rejecting demand requests is similarly problematic for UAM applications, given customer retention and wide-scale adoption will likely be strongly tied to reliability and availability of air taxis. Some strategies to increase reliability used in public transportation, such as a guaranteed ride home, may be valuable for UAM applications, e.g., if the UAM service cannot fly, the passengers would be given priority and guaranteed a ride via a ground transportation mode (like ride-hailing) for a similar or reduced price as UAM. Based on a survey of 2500 commuters in the U.S., Boddupalli et al. (2020) found that individuals were 1.8 times more likely to take an air taxi if a guaranteed ride home were provided.

All of these factors point to the trade-off between system performance and system cost—that is, we can over-design a system by ensuring extra aircraft in the fleet are available to serve peak periods and most customer demand requests, but serving all customer demand requests will likely be prohibitively costly. For example, the former CEO of NetJets, a private business jet company with fractional ownership, noted that in order to be profitable, he needed to cover 98 percent of all requested trips, and that serving 100 percent of all requested trips eliminated profits (Berger, 2001, as quoted in Mane and Crossley, 2007). Findings related to intelligent operational strategies and pricing policies that have been able to improve performance in ground transportation offer promising directions for the UAM community.

Optimizing over different time horizons will be important, particularly given the higher costs of establishing vertiports and charging stations for UAM applications than for EV applications. In addition, planning the deployment of vertiports and charging stations in ways that provide equitable access to citizens will be important if public funding is used for this infrastructure.