Study On Takeover Risk Of Automated Vehicles Considering Multiple Risk Sources

With the development of intelligent vehicles,the safety problems of automated vehicle’s takeover have become increasingly prominent.The existing research mostly focuses on the factors affecting takeover performance,and there is little quantitative research on the takeover risk of automated vehicles.Most of the existing studies use the collision time of the obstacles in front to describe the takeover risk.This method considers a single risk source and ignores the influence of other vehicle factors,road factors and driver factors,resulting in risk omission.Therefore,research on risk assessment of automated vehicle’s takeover considering the driver-vehicle-road factors has practical significance for comprehensively and accurately describing the takeover risk and improving the safety of automated vehicle’s takeover.Firstly,based on the driving simulator platform,the influence of personality traits and urgency on takeover performance is studied,at the same time,it is verified that there are some shortcomings in using collision time of obstacles in front to describe takeover risk.On this basis,this paper comprehensively considers the multiple risk sources,and introduces the field theory into the quantitative research of takeover risk.Aiming at the problem of overestimating and underestimating the risk,this paper improves the driving safety field by considering the shape of traffic physical objects,and then puts forward the takeover safety index.Finally,the takeover warning algorithm is proposed,and the effectiveness of the method is verified by experiments.The main contents are as follows:(1)Automated vehicle’s takeover can be affected by many factors.Most of the existing studies use the collision time to describe the risk of takeover.Therefore,based on the driving simulator experiment,this paper verifies the shortcomings of the method only considering a single risk.At the same time,aiming at the influence of individual differences of drivers on takeover performance,this paper analyzes the takeover time and quality of automated vehicle’s takeover considering the driver’s personality traits and takeover urgency.(2)Because of the shortcomings of the existing characterization methods of takeover risk,this paper comprehensively considers multiple risk sources,introduces the field theory into the quantitative study of takeover risk,and uses the distribution of field to simulate the distribution of risk.Driving Safety Field can comprehensively depict the risk of driver-vehicle-road factors,but it has the problem of overestimating and underestimating the risk around objects.Therefore,considering the influence of traffic physical object shape on the field distribution,this paper improves the existing Driving Safety Field.The improved potential energy field,kinetic energy field and behavior field are used to represent the risks of static objects,moving objects and drivers’ characteristics to the surrounding environment.On this basis,the takeover safety index is established,and a risk assessment method of automated vehicle’s takeover considering multiple risk sources is proposed.(3)In order to ensure the safety of takeover and improve the driver’s comfort.Based on the risk assessment method of automated vehicle’s takeover,the relative takeover safety index is established,and a human-machine friendly takeover warning algorithm is proposed.Aiming at the problem that it is difficult to quantify the driver’s risk factors,based on the influence of personality traits on takeover performance,this paper proposes a driver’s takeover risk factor suitable for takeover scenarios,and calibrates the influencing factors and parameters in the model.Finally,the effectiveness and superiority of the proposed takeover risk assessment method and takeover warning algorithm are verified by experiments.