Associating stated preferences of emerging mobility options among Gilbert City residents using Bayesian Networks

In the planning stage of emerging mobility options, residents' stated preferences are important to understand the acceptability, co-existence, demand estimation, and utilization. Thus, this study applied Bayesian Networks (BN) on the survey data from Gilbert City, Arizona, to explore the residents' stated preferences for Autonomous vehicles (AVs). It further explored the AVs association with four mobility options - electric vehicles (EVs), electric scooters, and docked and dockless bikes. It was revealed that about 66 % of respondents who want AVs would use them, making about 34 % overestimation when want is used to estimate demand. Furthermore, respondents interested in using EV charging stations are also more likely to want the AVs in the city as well as use them. No significant difference was observed between docked and dockless bikes on wanting the AVs but on using them. Furthermore, respondents who would not use dockless bikes have the highest predicted percentage difference between wants and use of AVs. The joint analysis of the variable revealed that the highest prediction of using AVs is attained for male respondents who would use all emerging mobility options. The practical application of this study is presented along with recommendations to operators, city engineers, and planners.

Automated summary

This study used Bayesian Networks to analyze the stated preferences of residents in Gilbert City, Arizona for Autonomous vehicles (AVs), and explored the association between AVs and other mobility options. The results showed that approximately 66% of respondents expressed a desire to use AVs. Respondents who were interested in using EV charging stations were more likely to want and use AVs. Among those who had differences between their intention and actual use of AVs, respondents who did not use dockless bikes had the largest percentage difference.