Research On The Impact Of Asphalt Pavement Rutting On The Lateral Safety Of Autonomous Vehicles

Autonomous driving,as a cutting-edge technology in the field of transportation,has received widespread attention worldwide,and the frequent auto-driving traffic accidents in recent years have also aroused deep concern for the safety of autonomous driving in the society.Rutting is one of the main diseases of asphalt pavement,which is very harmful to driving safety.However,most of the current research and development institutions have carried out a large number of researches on smart roads and vehicle-road coordination,but they have not paid enough attention to the impact of rutting on the lateral safety of autonomous vehicles.This paper aims to build a closed-loop simulation system for autonomous vehicles-rut roads to study the impact of rutting on the lateral safety of autonomous vehicles,and build a lateral safety risk prediction model for autonomous vehicles on rutted roads based on the impact of rutting to quantify autonomous driving The lateral safety risk of vehicles on rut roads.Based on literature research,the rutting analysis index based on dangerous driving conditions is determined from the two perspectives of rutting reduction of road unevenness and anti-skid performance.From three aspects of vehicle rollover,sideslip and yaw instability,dynamic evaluation indicators and their limits are selected to characterize the lateral safety of autonomous vehicles.Car Sim was used to build the vehicle subsystem and vehicle model.Considering the road plane alignment,rut geometry and road dry and wet conditions,the rut pavement simulation model was established.Based on preview-following theory and linear quadratic optimal control theory,a lateral controller was designed in MATLAB/Simulink environment,and lane change trajectory planning models of straight and curve were established to simulate the rut crossing of an autonomous driving vehicle.Then the autonomous vehicle-rutting road simulation system is built.The dynamic response of the autonomous vehicle crossing the rutted pavement in different scenarios is simulated and compared with that of the manual vehicle.The influences of rut width,rut depth and side wall slope on roll-over,sideslip and yaw instability of the autonomous driving vehicle at different speed levels are analyzed.The results show that,compared with manual driving,the automatic driving lateral control algorithm reduces the lateral safety risk of vehicles on rut roads.However,in terms of control accuracy,rutting significantly reduces the control effect of the automatic driving lateral control algorithm.It is still necessary to consider bad road conditions and improve the robustness of the control algorithm.After the rut is widened,the rollover risk of the vehicle is reduced overall,but when the width exceeds 700 mm,the possibility of yaw instability and sideslip increases.From the perspective of marginal effects,it is recommended that the rut width reaches 700 mm and be maintained in time.The deeper the rut,the higher the possibility of rollover,sideslip and yaw instability when the vehicle crosses the rut.For self-driving vehicles that change lanes at high speed on curved roads,even if the rut depth is only 5mm,they will face the risk of yaw instability and sideslip in the presence of water.The more the rut sidewall trembles,the higher the possibility of roll-over when crossing the rut.However,the influence of sidewall slope on sideslip and yaw stability of the autonomous vehicle is limited.Considering the influence of vehicle speed and curve radius,a lateral safety risk prediction model for autonomous vehicles crossing straight rut roads and curved rut roads is established based on the partial least square regression method.The auxiliary analysis results show that the model fitting error is small and the prediction accuracy is high.The auxiliary analysis results show that the model fitting error is small and the prediction accuracy is high,thereby quantifying the lateral safety risk of autonomous vehicles on rut roads.The vehicle speed and the radius of the circle are also significant influencing factors.According to the prediction model obtained by regression,the risk of rollover,sideslip and yaw instability of autonomous vehicles can be predicted when different rutting geometric indicators change.The research results of this paper lay a foundation for further revealing the relationship between rutting and lateral safety of autonomous driving vehicles in multiple scenarios,which can provide a reference for future safety studies of autonomous driving vehicles under complex road conditions,and have certain guiding significance for road maintenance,traffic safety assurance and control algorithm optimization under the background of autonomous driving.