Research On Driver Lane Change Depth Warning System Under High Speed Driving Environment

Driver lane change or overtaking is a complex process which must take into account both longitudinal and lateral movement of vehicles.Drivers are more likely to make improper judgments on the safety of the process,which has laid a hidden danger for traffic accidents.Advanced Driving Assistance Systems is a kind of vehicle active safety technology,which can enable drivers to obtain potential dangers in a very short time and give warning to improve driving safety.Aiming at the traffic problems caused by driver lane change in high-speed environment and the current development of lane change warning field,this paper studies the driver lane change depth warning system under high-speed driving environment.The following research has been done:(1)Predict collision time of dangerous driving under high-speed driving environment.Based on the principle of camera imaging,zhang zhengyou calibration algorithm was used to calibrate the camera,then the internal and external parameters of the camera were solved.The vehicle position and speed detection models at the rear and on the side were established to predict the collision time.(2)Detect potentially dangerous vehicles rear and on the side.Vehicles under high-speed environment were detected based on the deep warning network,then the data set required for training was made.Keras deep learning framework was used to train and test the network.The accuracy and timeliness of the deep warning network were tested by video in the evening,night and day.(3)Estimate the minimum safe lane change distance.In view of the complex behavior characteristics of the judgment lane change under high-speed driving environment,the geometric collision model of elliptical vehicle was established.Based on this model,the warning models of the influence of the target lane rear car,the target lane front car and the original lane front car on the lane change safety of the vehicle were established respectively.The established models are simulated in MATLAB respectively.(4)Experimental verification of lane change depth warning system.According to the demand of warning system,the platform of real vehicle experiment system was built.Then the longitudinal distance measurement,velocity measurement,lane change intention and lane change warning experiment were conducted,and the experimental results were analyzed.Based on the study of the driver lane change depth warning system under high-speed driving environment,this paper obtains an warning system that may accurately and quickly sense the motion state of surrounding vehicles and conforms to the driver’s habit lane change warning logic,which provides certain theoretical support and technical basis for the development of traffic safety and vehicle driving assistance technology.