Mode choice analysis of shared autonomous vehicles as first-mile service in Bangkok, Thailand

Abstract

Shared autonomous vehicles (SAVs) are a promising first-mile mode to access the public transportation. This study focuses on the mode choice analysis assuming that SAVs are introduced in Bangkok, Thailand. Stated preference experiments based on efficient design were conducted. The rescaled logit model was obtained from the nested logit trick on the combined revealed-preference and stated-preference data. The results showed that with the introduction of SAVs, the shares of existing taxi, motorcycle taxi (MC), and light public transit (LPT) modes will decrease by approximately 28.24%, 27.50% and 21.93%, respectively. These indicate that people who are accustomed to using taxi and MC as their first-mile services are respective most and second most likely to change their choices to SAVs. The values of times for income groups were estimated. Service providers should consider reducing travel cost to attract more lower-income users, and reducing travel time to attract more higher-income users. The average elasticities and simulation results imply that the improvement of the travel cost of SAVs should be a priority over travel time in order to effectively attract more individuals to the SAVs mode, and the improvement of travel time of SAVs leads to the most shift from taxi and MC.