Air Travelers' Preferences for Multimodal Urban Air Mobility (UAM) as an Airport Shuttle: A Stated Preference Study

Abstract

11 Urban Air Mobility (UAM) presents a promising solution to alleviate traffic congestion and reduce 12 travel time. However, UAM trips inherently require take-off and landing at dedicated vertiports, making 13 them multimodal by nature—a factor often overlooked in the existing literature. This study investigates 14 passengers' preferences for multimodal UAM transport compared to ground transport, with a specific 15 focus on how access modes to vertiports influence traveller attitudes in an airport access scenario. To 16 examine these preferences, a two-stage stated preference (SP) experiment was designed, and a nested 17 logit model was applied to analyse travellers’ intentions to choose UAM-integrated multimodal 18 transport options for airport access. The findings aim to provide researchers, policymakers, and 19 practitioners with a deeper understanding of public adoption of UAM services. Additionally, the results 20 are expected to inform the development of UAM networks and their integration with ground transport 21 systems. 22 Keywords: Urban Air Mobility, multimodal passenger transport, airport shuttle, stated-preference 23 survey

25 1 Introduction 26 As ground transport systems approach capacity limits, cities worldwide face growing traffic congestion. 27 According to INRIX (2025) which analysed congestion levels in 946 urban areas globally, 55% 28 experienced increased traffic delays compared to 2023. With workers returning to offices, this trend is 29 expected to worsen. In contrast to well-developed ground transport systems, low-altitude airspace 30 remains underutilised. Recent advancements in vertical take-off and landing (VTOL), battery 31 technologies, and automation have enabled the development of Urban Air Mobility (UAM). Utilising 32 electric vertical take-off and landing (eVTOL) aircraft, UAM offers a promising solution to urban 33 congestion by providing significantly faster travel speeds (can be 150-200 mph, see Holden & Goel 34 (2016)) and bypassing ground traffic altogether. 35 Due to its unique characteristics—such as low-altitude operation, fast travel speeds, and higher costs— 36 researchers have extensively examined potential users’ attitudes toward UAM. Examples include 37 Boddupalli et al. (2024), Coppola et al. (2024), Karimi et al. (2024) and Riza et al. (2024), which 38 conducted surveys to explore travellers’ expectations, concerns, and willingness to use UAM compared 39 to traditional modes (cars, public transit, and taxis) across scenarios such as commuting, intercity travel, 40 and airport access. These studies found that, compared to existing ground transport modes (especially 41 car), respondents’ willingness to use UAM is mixed, with some found negative (Boddupalli et al., 2024; 42 Fu et al., 2019; Jang et al., 2025) and others found positive (Cho & Kim, 2022; Coppola et al., 2024; 43 Samadzad et al., 2024). As UAM services are significantly more expensive than other options, high cost 44 has been identified as a major barrier to adoption (Cohen et al., 2021; Long et al., 2023; Straubinger et 45 al., 2020). Like many novel services and technologies, UAM primarily attracts young, well-educated, 46 and employed individuals (Brunelli et al., 2023; Fu et al., 2019; Song et al., 2024). Due to the high cost 47 of UAM services, studies have found that UAM adoption is generally higher among individuals with 48 higher incomes (Chae et al., 2024; Coppola et al., 2024; Karimi et al., 2024). 49 Unlike private cars or taxis, UAM operations rely on designated vertiports. In UAM’s early stages, the 50 vertiport network is likely to be sparse due to safety and noise concerns that require these facilities to 51 be located away from residential areas (Preis & Vazquez, 2022). As a result, vertiports may not always 52 be within walking distance, necessitating first/last-mile ground transport modes such as buses, shared 53 bikes, or taxis. Consequently, UAM trips are inherently multimodal. The convenience or inconvenience 54 of these access/egress modes can significantly influence travellers’ willingness to adopt UAM. Despite 55 advancements in UAM research, the impact of access/egress modes remains underexplored in the 56 existing literature. 57 To address this gap, this study employs a two-stage stated preference (SP) survey to examine the role 58 of ground transport access modes in UAM adoption. The airport access scenario serves as the study 59 context, given the suitability of UAM for long-distance, premium-priced trips and the higher income 60 levels of air travellers relative to the general population. In this survey, respondents first choose an 61 access mode and then select their primary travel mode to the airport. A nested logit model is used to 62 assess how access modes influence respondents’ choices. 63 The remainder of this paper is structured as follows: Section 2 reviews the relevant literature. Section 3 64 outlines the survey design and data collection methods. Section 4 details the modelling framework and 65 presents anticipated results. Finally, Section 5 discusses the conclusions and implications. 66 2 Literature review 67 The number of studies on Urban Air Mobility (UAM) has surged over the past five years. In 2018, 68 fewer than 80 publications existed, but this number nearly doubled to around 160 by 2020 (Abbasi et 69 al., 2024). UAM research spans a wide range of topics, including aircraft configurations, vertiport 70 design and location, public acceptance, demand estimation, and fleet operational planning (Garrow et 71 al., 2021; Rajendran & Srinivas, 2020; Sun et al., 2021). Among these, public acceptance and demand 72 estimation have garnered significant attention, as UAM, being a novel and capital-intensive transport 73 mode distinct from ground transport, requires widespread public support.

74 Current research on public acceptance and demand for UAM can be broadly categorised into two groups. 75 The first focuses on psychological factors, using technology acceptance models to analyse how 76 attributes like perceived safety and perceived benefits affect adoption intentions (Janotta & Hogreve, 77 2024; Karami et al., 2024; Vongvit et al., 2024). Among these factors, perceived safety is widely 78 recognised as positively correlated with UAM adoption. The second group examines the impact of trip 79 attributes (e.g., travel time, cost), personal characteristics (e.g., socioeconomic factors), and travel 80 habits (e.g., commonly used modes, air travel frequency). These studies frequently employ SP surveys 81 and discrete choice models, consistently identifying travel time and cost as key determinants. Notably, 82 they emphasise that reducing costs is essential for UAM to transition beyond a niche market (Asmer et 83 al., 2024; Rimjha et al., 2021; Wu & Zhang, 2021). 84 Vertiport accessibility is another major factor influencing UAM's success (Asmer et al., 2024; Rimjha 85 et al., 2021). Many studies use “access/egress time” as a proxy for accessibility, with assumptions 86 ranging from 5 to 20 minutes (Boddupalli et al., 2024; Coppola et al., 2024; Fu et al., 2019; Karimi et 87 al., 2024). A few studies further specify access modes: for example, Song et al. (2024) assumes an 88 average access time of 9 minutes by car or walking, Hwang & Hong (2023) assumes 4–14 minutes by 89 walking, and Rothfeld (2022) assumes 5–10 minutes on foot. However, vertiport location optimisation 90 studies present a contrasting picture. For instance, Rimjha et al. (2021) propose 50 to 200 vertiports 91 across 17 counties in Northern California, equating to only 9 × 10!" to 0.004 vertiports/km2. Similarly, 92 Rajendran & Zack (2019) assumed that only those who could save at least 40% in travel time by using 93 UAM would switch to this mode. Using New York trip data, recommend only 21 vertiports for New 94 York City (0.017 vertiports/km2). Asmer et al. (2024), referencing to Mayakonda et al. (2020)’s 95 assumption on vertiport density (0.002-0.007 vertiports/km2), assumed that vertiport density would 96 range from 0.001 to 0.002 vertiports/km2 in 2030 and 0.01-0.02 vertiports/km2 in 2050. This translates 97 to access/egress distances of 9-12 km in 2030 and 3-5 km in 2050. 98 The mismatch between vertiport density assumptions and accessibility modelling highlights the 99 unrealistic nature of assuming all vertiports are reachable by walking. UAM trips should be treated as 100 air-ground multimodal journeys, where the convenience of the access/egress stage significantly impacts 101 adoption decisions. While the effect of accessibility on travel demand has been well studied for both 102 ground transport (shared mobility and public transport) (Albrecht et al., 2025; Berg Wincent et al., 2023; 103 Chowdhury et al., 2016; van Soest et al., 2020) and air travel (Choo et al., 2013; Gupta et al., 2008; 104 Hess & Polak, 2006), the integration of UAM with ground transport has received less attention. Given 105 UAM’s unique access requirements and respondents' unfamiliarity with the mode (which may increase 106 perceived risk), the influence of the access/egress stages on adoption and travellers’ perceptions of its 107 multimodal nature remains underexplored. 108 3 Data 109 Since UAM is still a new concept and few travellers have experienced this service, we use a SP 110 experiment to collect data on respondents' travel mode choices. Given that UAM is more suitable for 111 longer intracity trips, we focus on airport access as the scenario. 112 The hypothetical trips to the airport are shown in Figure 1. For car and taxi (including ridehailing), the 113 trips are point-to-point, involving just one leg. For UAM and tube trips, since the distances to vertiports 114 or tube stations may exceed walking distance, respondents can use a taxi or bus to reach the station. We 115 assume the airport is directly connected to the UAM or tube system, so the egress distance is set to zero. 116 Thus, both UAM and tube trips are divided into two stages: the access stage and the main stage.

(a) Access airport by car or taxi

(b) Access airport by tube

(c) Access airport by UAM 118 Figure 1 Journey to the airport by various modes

119 The overall experiment consists of three parts: (1) Blocking questions to determine if respondents have 120 air travel experience. Those without such experience are screened out. If respondents are not screened 121 out, they are asked whether they have used air travel for business purposes and whether the car option 122 applies. These two questions serve to tailor the survey to respondents' real travel experiences (as show 123 in Table 1); (2) Eight stated mode choice scenarios, which form the core of the survey; and (3) Questions 124 on socio-demographic details, travel habits, and attitudes toward UAM. 125 Table 1 The relationship between travel experience and blocks

Block Used air travel for business Have a car that drives Block feature No. purposes regularly Business travel choice Car option in the choice scenario scenario Block 1 Yes Yes Yes Yes Block 2 Yes No Yes No Block 3 No Yes No Yes Block 4 No No No No

127 The choice scenarios are divided into two stages. First, respondents are asked to imagine a trip to the 128 airport and consider whether they would use a tube or UAM, taking into account walking time, in129 vehicle time, waiting time, and travel costs for the access modes. Based on these factors, they must 130 decide whether to stick with walking or opt for a bus or taxi as the access mode. In the second stage, 131 respondents choose their travel mode for the main journey, based on the two access modes they selected 132 for the tube or UAM in the first stage. If respondents feel that the selected access modes are unsuitable 133 for the second stage, they can always return to adjust their choices. The choice scenarios for both stages 134 are presented in Figure 2 and Figure 3.