Study On Mode Choice Behavior Of Autonomous Vehicle Based On An Extended Theory Of Planned Behavior

The rapid advancement of autonomous vehicle(AV)technology will have a revolutionary impact on the transportation system.The advent of autonomous vehicle technology will also bring an innovative on-demand motorized travel mode-shared autonomous vehicle(SAV).Past research has shown that not all new technologies are immediately welcomed by the public.AVs would have to likely go through a similar phase and would need to overcome not just technological challenges but also social barriers for successful penetration into the marketplace.Most previous studies on mode choice of autonomous vehicle focus on the analysis of objective attribute variables,which fails to extract deeper insights on travelers' intrinsic preference and potential influencing factors.This paper collected recent papers from databases such as Scopus,Web of Science,ScienceDirect and CNKI,and made a systematic review based on the perspective of influencing factors and research methods.On the basis of existing literature review,the key factors influencing travelers' intention of using AVs and SAVs were discussed by introducing knowledge and perceived risk as two predictive variables to expand the theory of planned behavior.Based on the five hypotheses established by deductive reasoning,a questionnaire was designed for this study.A total of 558 valid survey sample data were collected through online and field surveys.Statistical methods such as factor analysis,reliability test,validity test,and multiple regression analysis were used to conduct empirical research and analysis based on the survey samples.From the perspective of micro-users,the stated preference(SP)choice test method was used to investigate the influencing factors of travelers' mode choice of AVs and SAVs,and the conditional Logit model was established to specifically study the traveler's preference for AVs and SAVs,as well as the key factors affecting their choice intentions.From the analysis results of the model,it can be seen that personal attribute variables and travel attribute variables will affect the choice of AVs and SAVs to varying degrees.The multiple indicators and multiple causes model(MIMIC)and conditional Logit model with latent variables are established by integrating three categories of variables,including individual demographic variables,choice preference variables,and latent variables.This study not only verifies the applicability of the theory of planned behavior,but also further explores the relationship between usage intention among attitudes,subjective norms,perceived behavior control,knowledge,and perceived risks.Empirical results show that the discrete choice model integrating latent variables has better model fitting and stronger explanatory power.The model analysis shows that various latent variables in the framework of proposed model have different effects on mode choice of AVs and SAVs.The findings enhance the understanding of traveler mode choice motivation from psychological and service perspectives,providing data support for governments and companies in improving travel management strategies and product services in the future.