Research On The Effect Of Driver Trust Matching On Driving Control Take-over

Based on the development status and rights and responsibilities disputes of humancomputer driving intelligent vehicles,this study,from the perspective of human factors,considers the matching degree of the driver’s trust in L3 autonomous vehicle and the objective credibility of the system(hereinafter referred to as: trust matching),and the impact on the driver’s takeover of the vehicle.Based on this,this article has identified three research objectives:(1)To explore the impact of dynamic changes in driver trust on takeover decisions in dangerous driving scenarios of human-machine driving(hereinafter referred to as "dangerous human-machine driving"),and to clarify that the mismatch between driver trust and system objective credibility is an important factor leading to differences in takeover decisions.(2)Explore the objective reliability standard of Automated Lane Keeping Systems(ALKS)based on L3 autonomous vehicle,and realize the matching measurement of driver confidence based on ALKS.(3)Explore the impact of driver trust matching on take-over behavior and driving behavior,and analyze the proportion of control allocation after the driver takes over the vehicle.For the first research objective,video experimental testing was conducted.Firstly,the online platform conducted research and collected videos of dangerous human-machine driving from 6 different responsible parties.Subsequently,24 young drivers were recruited and divided into two test groups: questioning and trust through the "Driver Initial Trust Questionnaire" for experimental testing.The research shows that there is a big difference between the driver’s trust degree and the actual trust degree of the auto drive system in the risky machine co driving,which will affect the driver’s takeover decision.In the performance of the takeover decision,the test group questioned that drivers would choose to take-over the vehicle after 78% of the driving videos were played;After 68% of the driving videos were played,the trust test group drivers would choose to take over the vehicle,even though the auto drive system(ADS)in the video did not cause a traffic accident.For the second research objective,online research and expert scoring were used.Firstly,the auxiliary driving function standards of L3 autonomous vehicle ALKS were sorted out and analyzed,and an objective reliability evaluation questionnaire based on 10 auxiliary driving functions of ALKS was formed.Subsequently,four experts in the field of human factors were invited to rate the questionnaire.The final objective credibility standard based on L3 level autonomous driving ALKS was determined: a questionnaire trust score of less than 140 indicates insufficient trust,a score between 140 and 160 indicates moderate trust,and a score above 160 indicates excessive trust.Finally,the drivers who had previously participated in video experiments were contacted to fill out the questionnaire and compare it with the objective credibility standards formed by expert scoring.Finally,20 participants were divided into three test groups: insufficient trust,moderate trust,and excessive trust.For the third research objective,a simulation control take-over experiment was conducted.Build 12 dangerous human-machine driving scenarios through a simulation experimental platform to explore the differences in takeover and driving behaviors among the 20 test drivers with different trust matching situations.The results showed that the participants in the undertrust and over-trust groups took over more times than those in the moderate-trust group,and the over-trust group had a longer takeover time;The stability of lateral control and vertical control in the over-trust group is inferior to that of drivers in the under-trust and moderate-trust groups.In addition,the result of the driver’s control right distribution proportion shows that the driver believes that about 55% of the control right of the auto drive system is a reasonable and safe power distribution proportion in driving tasks such as emergency braking and steering.This study explores the impact of driver trust matching on take-over in risky partner machine driving;At the same time,an objective credibility standard based on L3 level ALKS was proposed to provide research ideas for measuring driver trust matching;Finally,a control allocation ratio based on the driver’s perspective was proposed,providing ergonomic suggestions for the safety design of control allocation in the co-driving of dangerous vehicles.