Study And Analysis On Mechanism Model Of Pedestrian Crossing Stop Intention Under Autonomous Driving

Autonomous driving has become an inevitable trend in the development of future traffic.However,in the study of autonomous driving,the research of domestic and foreign focused on the self-driving vehicle and related technologies,while the behavior under autonomous driving of other main players in the road traffic system,pedestrians,there are few feature studies.In view of the shortcomings of the research on the unmanned environment,this paper studies the pedestrian behavior mechanism with the vehicle unmanned as the background:Firstly,the function and operational logic of the VTPCDS(Vehicle-to-Pedestrian Communication Display System)is designed.In the driverless environment,the pedestrian can not communicate with the driver,but the vehicle can send messages to the pedestrian through the communication device.The existing solutions of vehicle-to-pedestrian communication have disadvantages in content and mode,what's more,the operation logic is unclear.This paper combines the development of detection and identification technology,analyzes the possible content and feasible mode of the VTPCDS,then the function and operational logic of the the VTPCD are designed as the technical basis for this article.Secondly,the mechanism model of pedestrian crossing stop intention under autonomous driving is constructed.On the basis of summarizing the research of individual behavior mechanism,this paper proposes the modeling ideas applicable to background of this research,then analyzes and extracts the potential influencing factors of the pedestrian behavior under autonomous driving,as the constituent elements of the mechanism model.After that,this paper establishes the significant impact hypotheses among the elements.Then,establishing the mechanism model of pedestrian crossing stop intention under autonomous driving as the theoretical basis for this article,of which personal attribution and road environment are inputs,pedestrian acceptance,perceived usefulness and ease of use are intermediate variables,and crossing stop intention is output.Thirdly,the mechanism model hypothesis test and influencing factors analysis based on experimental data.Considering the unpopularity of autonomous driving and the need for sample size for hypothesis testing,this paper collected 3216 valid experimental data by using the SP questionnaire.According to the relevant assumptions between the model model elements and the data types,multiple linear regression or Logit models are selected for confirmatory modeling,aiming at hypothesis testing and model calibration.In this paper,the multiple linear regression model of pedestrian acceptance,the BL probability model of feedback information perceived usefulness,the MNL probability model of feedback class information perceived ease of use,the Nested-Logit probability model of the usefulness of auxiliary information perceived usefulness and The BL probability model of the pedestrian crossing stop intention.The likelihood ratio test and the goodness of fit index indicate that all these models are valid,and the significant impact hypotheses of the mechanism model are partially established or fully established.According to the model calibration results,the influencing factors are analyzed,and the mechanism model is corrected according to the hypothesis test results.Finally,the heterogeneous effect of the influencing factors in the mechanism model is studied.The trend of sample population in the traditional Logit model often masks the individual's preferences.To improve the robustness of the mechanism model and to study the differences of pedestrians' preferences,Latent Logit model is used to study the heterogeneous effect of the influencing factors in the mechanism model.The results show that there are heterogeneous effects of some influencing factors,and trying to explore its causes.