Urban aerial mobility: Network structure, transportation benefits, and Sino-US comparison

Abstract

Large cities worldwide suffer from traffic congestion, leading to huge economic, social, and environmental costs.1Zhang L. Zeng G. Li D. et al.Scale-free resilience of real traffic jams.Proc. Natl. Acad. Sci. USA. 2019; 116: 8673-8678Crossref PubMed Scopus (84) Google Scholar Various measures have been taken, spanning congestion pricing, public transportation promotion, high-occupancy driving lanes, and road space rationing to recent on-demand ride-sharing models, etc.2Hanna R. Kreindler G. Olken B.A. Citywide effects of high-occupancy vehicle restrictions: evidence from “three-in-one” in Jakarta.Science. 2017; 357: 89-93Crossref PubMed Scopus (32) Google Scholar,3Vazifeh M.M. Santi P. Resta G. et al.Addressing the minimum fleet problem in on-demand urban mobility.Nature. 2018; 557: 534-538Crossref PubMed Scopus (211) Google Scholar,4Eliasson J. Efficient transport pricing–why, what, and when?.Commun. Trans. Res. 2021; 1: 100006Crossref Scopus (24) Google Scholar However, none of them can solve this inherent problem due to the asymmetry between transport infrastructure and dynamic travel demand. Opportunities to expand road capacity are rare, but an emerging on-demand transportation mode, urban aerial mobility (UAM), is unlocking tremendous traffic capacity in the urban low-altitude space, leading to a perfect solution to traffic congestion.5McKinsey & CompanyThe future of air mobility: electric aircraft and flying taxis.https://www.mckinsey.com/featured-insights/the-next-normal/air-taxisDate: 2022Google Scholar The rise of UAM is thanks to the recent developments of electric vertical-takeoff-and-landing (eVTOL) vehicles. These vehicles leverage the infrastructure of vertiports for takeoff and landing and are coupled with ground vehicles to provide commuting services, where ground vehicles offer first-mile and last-mile travels to and from vertiports, and eVTOLs are used to commute between vertiports.6Kai W. Jacquillat A. Vaze V. Vertiport planning for urban aerial mobility: an adaptive discretization approach.Manuf. Serv. Oper. Manag. 2022; 24: 3215-3235Crossref Scopus (1) Google Scholar The whole UAM industry is expecting a bright future, but there is a lack of systematic understanding of it. Specifically, what would be the network structure of UAM in cities? What are the transportation benefits of UAM? How does UAM affect traffic congestion? These questions will be fundamental for stakeholders to invest in UAM, for city authorities to promote UAM as a new commuting option, for regulation makers and UAM operators to plan for the deployment, etc. To fully understand UAM, the first step is to deploy the vertiport network since it constitutes the foundation of UAM, and the next is to investigate UAM operations and the potential demands for interpreting the transportation benefits of UAM. Specifically, inside a metropolitan area, how do we select the locations of vertiports to support UAM operations? Built upon a UAM network, what would be the interactions between the supply side of UAM operations and the demand side of passengers’ adoption? Once we have the UAM network and the interactions between supply and demand, we can proceed to interpret the UAM system. Here, we leverage the integrated UAM planning model proposed by Wang et al.6Kai W. Jacquillat A. Vaze V. Vertiport planning for urban aerial mobility: an adaptive discretization approach.Manuf. Serv. Oper. Manag. 2022; 24: 3215-3235Crossref Scopus (1) Google Scholar We investigate the UAM systems of major metropolitan areas or cities in the US and China to have a systematic understanding of UAM and to make comparisons. At a city or a metropolitan area level, the physical existence of a UAM system lies in the network of vertiports that constitutes the foundation of UAM operations. Figure 1A shows the networks of four representative metropolitan areas or cities in the US and China, respectively. In general, a UAM network is a hub-and-spoke network structure where flying by eVTOLs is for connections between hubs (i.e., vertiports) and ground transportation is for connections between spokes and hubs. However, compared with other traditional hub-and-spoke networks, e.g., airline networks, the main difference is that the hubs are not used to aggregate demand to operate large vehicles/aircraft to lower the cost per passenger, and they are used to enable more frequent flights between vertiports, and thus more frequent eVTOL movements, to improve the level of service reliability of UAM. This essentially attributes to the difference between the on-demand business model and the on-schedule business model, where the former emphasizes service reliability, and the latter emphasizes capacity utilization. A UAM system is centralized for a city or metropolitan area and covers urban areas instead of being widely spread over urban areas and rural areas. The reason is that a centralized system enables efficient operations that can, on the one hand, reduce operational costs and, on the other hand, improve the level of service to attract more customers. UAM provides a better option for medium-to long-distance commutes. This is due to the fact that for a UAM trip, there will be detours for the first-mile and last-mile transportation, by using the ground transportation. For short-distance commutes, the benefit of a short flight cannot compensate for the loss in the detours and thus does not lead to an acceptable travel time reduction of UAM compared with direct door-to-door ground transportation. This indicates that UAM will be more like a substitute for commuter rails (usually for medium-to long-distance commutes) as opposed to local bus systems (usually for short-distance commutes). UAM systems are varied across different cities. In the New York Bay Area, San Francisco Bay Area, and the Guangdong-Hong Kong-Macau Greater Bay Area, the UAM systems are constructed to bridge these bay or water areas for better connectivity. For cities with better road connections and dominated downtown areas, UAM systems are constructed to better connect business districts, e.g., Beijing, Washington, DC, etc. There is a fundamental difference in UAM systems between the US and China. A UAM system in the US normally spans a metropolitan area of multiple cities, e.g., cities in the San Francisco Bay Area, and thus better integrates the area. However, a UAM network in China is typically associated with a mega-city, e.g., Beijing or Guangzhou, and makes the city more centralized. The reason is that cities in China are relatively larger and denser in both population and area and thus need dedicated UAM systems, but cities in the US are relatively small and normally affiliated with a metropolitan area and thus need shared UAM systems. To get insights into the transportation benefits of UAM, we now dive into the patterns of UAM trips. The advantages of UAM come mainly from flying (Figures 2A and 2B ) since it provides transportation over most of the travel distance of a trip but does not take the most travel time. Flying times are short, several minutes, which is due to UAM systems being centralized. This suggests that UAM is not so restricted by the battery technology of eVTOLs for urban deployment. First mile and last mile, relying on ground transportation, are the major issues and consist of most UAM trip times. Comparatively, the first- and last-mile travel time is much longer in China compared with that in the US. The reason is that cities in China are much more congested. We now verify the benefits of traveling by UAM compared with traveling by ground transportation. When the ground travel time is longer, the associated UAM trip will save more time (Figures 2C and 2D) since more benefits can be obtained by flying over a long distance. This echoes our previous finding that UAM is mainly for medium- to long-distance commutes. On average, UAM reduces the origin-destination (O-D) travel times by 30%–40% and 40%–50% for major metropolitan areas or cities in the US and China, respectively (Figures 2E and 2F). It is interesting to see that UAM turns out to be more beneficial for cities in the US than those in China. The main reason is that a UAM trip consists of first- and last-mile travel and flying, and for cities in China, the first- and last-mile travels are near the areas of vertiports, where the ground transportation is already congested and thus limits the benefits of UAM trips in China. Figure 1B shows the traffic changes in representative cities in the US and China after using UAM. For main roads or highways that connect medium- to long-distance travels, the traffic volume drops significantly, and this is reasonable because UAM absorbs a large proportion of medium- to long-distance commutes, which will remarkably alleviate the congestion thereof. However, for minor roads that connect locally, the traffic volume rises slightly, and the reason is that the first- and last-mile travels in the UAM are near vertiports and thus will attract more ground traffic locally, which makes the local areas more congested. Therefore, we can conclude that UAM can solve the congestion problems for main roads or highways in cities or metropolitan areas, which are indeed growing and critical problems for urban transportation. However, UAM, based on current air-flying-only eVTOLs, does make the local areas more congested, which deserves more attention before pushing UAM to the front. We need to better improve the accessibility to vertiports in local areas and reduce the local ground traffic; otherwise, we will create further trouble for urban transportation. For example, we can re-design bus lines to better serve local areas, and we may design shared-bus services to provide more convenient and efficient local transportation options. Alternatively, we may push eVTOL technologies further such that eVTOLs will not need to leverage the infrastructure of vertiports to take off and land. Henceforth, they can provide door-to-door flying services, and that will solve the first-and-last-mile problem once and for all. We leverage the mathematical programming model in Wang et al. to optimize the UAM networks, which yields all outputs for the above analysis.6Kai W. Jacquillat A. Vaze V. Vertiport planning for urban aerial mobility: an adaptive discretization approach.Manuf. Serv. Oper. Manag. 2022; 24: 3215-3235Crossref Scopus (1) Google Scholar Here is a brief description of the model: we first select a set of candidate locations for building vertiports. The main decisions are whether to build a vertiport at each candidate’s location or not, which will result in a UAM network. However, these decisions do not stand alone since there are strong interdependencies between this strategic vertiport deployment and future UAM operations. In addition, the model also needs to consider the impacts of UAM operations on the passengers’ preference for using UAM for commuting. In all, the vertiport deployment is strategic planning, built upon which there is a feedback loop between UAM operations and passengers’ demands of UAM, and the objective of the planning model is to maximize the daily profit. Details of modeling the UAM operations and passengers’ demands are also referred to in Wang et al. The main input data of the model are the travel demand data and travel time. For the US, we resort to the Longitudinal Employer-Household Dynamics (LEHD) Origin-Destination Employment Statistics (LODES) database from the US Census Bureau (2022), which estimates, for each census block pair, the number of people residing in one and working in the other.7U.S. Census BureauLongitudinal Employer-Household Dynamics. LEHD, 2022https://lehd.ces.census.gov/daGoogle Scholar We define travel demand in the 10 largest US metropolitan areas during the morning rush hours accordingly and aggregate it at the zip code region level. For China, high-resolution demand data in the 12 largest cities are extracted from hand-phone signaling data. For aggregation, each city is divided into a grid with 0.025° (longitude) and 0.025° (latitude) regions, and the trips are mapped into regions in the grid. The travel demand considered between a pair of regions is the number of trips during the morning commute period. The ground travel time for each pair of origin and destination is imputed by using Google API and Amap API for the US and China, respectively. All other parameters come from Holden and Goel (2016).8Holden, J., Goel, N. (2016). Fast-forwarding to a future of on-demand urban air transportation. San Francisco, CA.Google Scholar This work was supported by the National Natural Science Foundation of China (NSFC) under grant numbers 52221005 and 52220105001. The authors declare no competing interests.

Topics & Concepts