A Study On Travel Mode Selection Of Autonomous Vehicles Based On SEM-NL Model

With the rise of autonomous vehicles,automotive companies are actively developing,testing and marketing autonomous vehicles,and autonomous vehicles have been put into operation in many cities at home and abroad.Autonomous vehicles not only enable travelers to engage in a number of productive or recreational new activities while traveling,providing greater comfort and convenience for travelers,but they also significantly improve the passenger experience and offer significant benefits in terms of safety,maneuverability as well as economy.However,with the rapid adoption of autonomous vehicles,urban residents have more choices for their addresses and travel methods.Therefore,this thesis introduces autonomous vehicles into the transportation system,and explores the impact of autonomous vehicles on the joint choice of residential address and travel mode.Based on a comprehensive analysis of the mode of travel of autonomous vehicles,this thesis mainly scrutinizes and assessments the relevant literature on residential address and travel mode selection at home and abroad,and summarizes the influencing factors and research methods of residential address choice and travel mode selection.Secondly,a structural equation model of residents’ acceptance of autonomous vehicles is proposed,in order to ensure the validity and reliability of the questionnaire data,the reliability and validity of the sample data were also checked using Stata software,and lay a foundation for the next step of model analysis.Thirdly,through the survey method design of the psychological latent variables of the choice preference of autonomous vehicles,the sample data of residents’ joint choice preference is studied by combining revealed preference and stated preference.Finally,taking Wuhan as an example,an empirical analysis is carried out,and the NL model and SEM-NL model are established to evaluate and explore the effect of autonomous vehicles on the choice of place of residence and overall travel modes,and propose strategies and recommendations to improve people’s willingness to choose autonomous vehicles.This thesis provides a new analytical model of co-choice behavior,finding that the SEM-NL model,which considers the stimulus of inner embryonic variables,has more accurate explanatory power than the conventional NL model and also predicts co-choice more accurately.At the same time,this thesis proves that psychological latent variables have a non-negligible influence on residents’ joint choice behavior,and concludes the difference in the influence of autonomous vehicles on residents’ choice intention on different psychological latent variables.Under the hypothetical scenario,the impact of autonomous vehicle travel cost,travel time,walking and waiting time on residents’ choice is quantified through elastic analysis,which provides certain data support for governments and enterprises to formulate autonomous vehicle travel management strategies in the future.