Research On Ride-sharing Behavior Of Shared Autonomous Vehicles

With the development of autonomous driving technology and the spread of sharing economy,shared autonomous vehicles(SAV)have made shared mobility possible.This has significant practical implications for alleviating problems such as traffic congestion and environmental pollution caused by traditional modes of transportation.However,current research on SAV sharing overlooks consumer heterogeneity and psychological latent variables.Therefore,this study aims to consider consumer heterogeneity,elucidate the influence of psychological latent variables and individual attributes on consumer willingness to share SAVs.To explore the factors influencing consumer willingness to share SAVs,this study uses the extended theory of planned behavior and the technology acceptance model as a theoretical framework,designs a survey questionnaire,and collects 568 valid responses.Based on latent class modeling,consumers are segmented and a structural equation model is constructed to analyze the factors and pathways influencing heterogeneity in consumer willingness to share SAVs,in order to identify key factors affecting this willingness.The main conclusions are as follows:(1)The existence of consumer heterogeneity in SAV sharing is demonstrated.Based on latent class modeling using valid psychological latent variable data in the survey,the total sample is divided into three classes,with each evaluation index achieving the best comprehensive performance and a classification accuracy of 96.45%,demonstrating the validity of the model.The three classes are named positive user type,conflicted user type,and conservative rejection type,accounting for 20.42%,41.37%,and 38.21% respectively,based on their conditional probability and characteristics.(2)A structural equation model for SAV sharing willingness that considers consumer heterogeneity is constructed,and the influence relationships among psychological latent variables are analyzed.Based on consumer heterogeneity,three structural equation models are built,with explanatory power of 0.85,0.50,and 0.71 respectively,indicating that the latent variables explain willingness to share SAVs well and the model is effective.The study finds that attitudes,trust,perceived usefulness,and government regulation have a positive impact on the willingness to share SAVs for all three consumer classes,while perceived behavioral control is not significant for conflicted user type and conservative rejection type,and perceived ease of use is not significant for positive user type.Perceived risk is not significant for any of the three latent classes.(3)A multiple-causation and multiple-causality model was established to explore the impact relationships between consumer personal attributes and psychological latent variables of three potential subgroups.The study found that for positively inclined consumers,the frequency of shared car usage and the size of the city directly influence the consumer’s willingness to carpool,while other factors indirectly affect their willingness to carpool.For consumers with conflicting attitudes towards carpooling,the frequency of taxi usage,the size of the city,annual income,age,and education level directly affect their willingness to carpool,while other factors indirectly affect their willingness to carpool.For conservative and rejecting consumers,parking fees,daily travel distance,age,gender,and education level directly affect their willingness to carpool,while other factors indirectly affect their willingness to carpool.The research results verified the applicability of the combined method of latent class model and structural equation model in the field of SAV ride-sharing willingness research.The study has revealed the underlying mechanisms of SAV ride-sharing intentions for different latent classes,providing insights for governments and businesses to formulate regulatory policies and optimization recommendations,and to promote SAV ride-sharing in the future.