Citation

Wali B, Santi P, Ratti C. Transp. Res. C Emerg. Technol. 2021; 129: e103276.

Copyright

(Copyright © 2021, Elsevier Publishing)

DOI

10.1016/j.trc.2021.103276

PMID

unavailable

Abstract

An adequate understanding of consumer affinity towards automated vehicles (AVs) is necessary to more reliably forecast the potential impacts of AVs on transportation system. The road to fully automated driving systems pass through partially automated driving systems. This study contributes by simultaneously analyzing consumer affinity towards partially (Level 3/4 automation) connected automated vehicles (CAVs) and fully AVs (FAVs) (Level 5 automation), and how it relates to sustainable travel behavior, user valuation of different vehicle attributes, (electric) vehicle ownership, and safety concerns. Unique data for over 3500 households from the 2017 California Vehicle Survey are harnessed. With an on-average positive correlation between consumer affinity towards partially CAVs and FAVs (Kendall's τb of 0.506), significant contrasts are also observed. Compared to around 64% of the households with positive affinity towards partially CAVs, only 35% were willing to consider buying a FAV in future. Around 30% of the households had negative affinity for FAVs but positive affinity for partially CAVs. A comprehensive simulation-assisted joint bivariate ordered discrete outcome modeling framework is used, considering preference heterogeneity at different geographic/ecological levels, from the region to household level, and further within and across the choices for partially CAVs and FAVs. Participation in carshare, rideshare programs, and more transit trips per member were independently associated with a greater affinity towards partially CAVs and FAVs. Ownership of a plug-in electric vehicle was associated with a 19.9% and 9.6% increase in the likelihood of highest affinity levels towards partially CAVs and FAVs, respectively. Some evidence is found that consumers may possibly use shared AVs, but not necessarily in combination with ride sharing. Simultaneous correlations between affinity towards partially CAVs and FAVs and most valued vehicle attributes, safety concerns, and sociodemographic factors are quantified.

METHODologically, the results underscore the importance of considering preference heterogeneity at different geographic/ecological levels. Captured through random parameters, the associations of several factors with consumer affinity towards partially CAVs and FAVs varied in magnitude and direction. In addition, statistically significant interactive effects (opposite and counterbalancing) of the latent unobserved factors underlying several random parameters were observed. The presence of multilayered preference heterogeneity highlights the personalized and complex nature of behavioral phenomenon underlying user acceptance of AVs. By reflecting the users' preferences towards CAVs and FAVs and characterizing the taste/preference heterogeneity at multiple levels, the findings can be integrated in larger travel forecasting models to more accurately predict the future-use and ownership patterns of CAVs and FAVs by various sociodemographic cohorts.

Language: en

Keywords

Automated vehicles; Bivariate random parameter discrete choice model; Interactive heterogeneity; Partially and fully automated vehicles; Preference heterogeneity; User preferences