Research On The Strategy Of Switching Longitudinal Driving Rights In The State Of Cognitive Distraction

As China’s comprehensive national strength continues to increase,the automobile industry,as one of China’s pillar industries,has ushered in a period of rapid growth and vigorous development,and automobiles have entered every ordinary family as "rigid" consumer goods.On the one hand,cars make our travel more convenient and faster,but on the other hand,it inevitably brings unsafe factors.At present,the number of people who die in traffic accidents in China reaches more than 63,000 people every year,and the traffic safety situation is still grim.In the road traffic system composed of people,vehicles,and road traffic environment,the driver is the weakest link in the system.Automobile intelligence provides an effective means for solving traffic safety accidents.However,in the process of moving from driving assistance to highly automated driving,the car will be in the stage of human-machine cooperative driving for a long time.Therefore,study the internal mechanism between the driving state,driving ability and safe driving of human drivers,explore the driving right switching strategy based on the driving state,deeply understand the driving behavior characteristics of the driver,and realize the codriving of human-machine cooperation of automatic driving,It has important theoretical significance to improve the safety of autonomous driving.This article aims at the problem of mandatory switching of driving rights from driver to vehicle in the case of man-machine co-driving.Starting from the state of the driver and driving performance,the physiological characteristics and driving behavior of the driver in the bad state of cognitive distraction characteristic.Revealing the changing rules of the driver’s control ability of the vehicle under different levels of cognitive distraction,quantifying the driver’s driving ability in real time according to the vehicle’s motion state,and proposing a longitudinal driving right switching strategy based on the driving ability failure interval for driving in the co-driver mode Power switching provides new methods.The main content of the paper is as follows:1)Based on UE4 software and logic G29 hardware,using Carsim vehicle dynamics model,a driver-in-the-loop simulation platform was built,and a car-following driving experiment scenario was established;A typical dual-task experiment in human factors engineering was carried out.By introducing the N-Back cognitive distraction subtask,the influence of cognitive distraction on driving safety was studied.Studies have shown that the time domain characteristics,frequency domain characteristics,and skin electrical signals of the driver’s HRV all have a clear linear trend with the degree of cognitive distraction.The time domain characteristics of HRV are most sensitive to the degree of cognitive distraction;Cognitive distraction has a significant effect on the longitudinal motion state of the vehicle,satisfies the Yerkes-Dodson rule,and has a certain effect on the lateral stability of the vehicle,but has no obvious effect on the driver ’s lane keeping ability;As the degree of cognitive distraction increases,the driver’s lane departure risk and rear-end collision risk increase.2)The Gaussian mixture model(GMM)and KL divergence methods are used to cluster the driving styles.The entropy weight method is used to calculate the ability weight of various drivers.The sliding window algorithm is used to calculate the driver’s offline driving ability,and a BP neural network model for real-time evaluation of driving ability is established.Studies have shown that the identification of driving ability based on the vehicle’s motion state can effectively characterize the driver’s driving behavior during the vehicle’s driving process,and quantify the driving ability change law under different cognitive distraction states.3)The driving ability failure interval is defined,the threshold calculation and realtime updating method of driving ability failure interval based on the distribution characteristics are proposed,and the best window parameters of the sliding window algorithm for calculating the optimal driving ability are studied.The research shows that the optimal time window length is 5s,and the time window coincidence is 4s.Under this parameter,the detection rate of the longitudinal tail collision accident by the algorithm is100%,and the average decision advance time is 5.6s.The driving right switching strategy based on the ability failure interval provides a new scientific method for longitudinal driving right switching in the autopilot man-machine co-driving mode.