Left-turn Assisted Collision Warning Algorithm Based On V2X

Transportation has gradually become a necessary part of People’s Daily life.Frequent traffic accidents in cities are threatening the safety of people’s lives and property.A large proportion of traffic accidents occur at intersections.As the connection area of multiple roads,the road conditions are more complex,and the left turn of vehicles is the key to affect the traffic efficiency of the intersection.Therefore,the left turn assistance can not only reduce the occurrence of accidents but also improve the traffic efficiency.Compared with the warning algorithm of single road driving,the warning algorithm of intersection is more complex,the behavior of vehicles has more possibilities,and the traffic flow is complicated and changes in real time.Therefore,the algorithm of intersection needs higher computational amount and real time.Advanced auxiliary driving generally adopts laser radar,millimeter wave radar and camera equipment to perceive the environment,the obstacles that affect perception equipment blocked out,causing recognition error,but the urban intersection is easy to appear obstacles,V2 X technology can breakthrough the limitation of the obstacles to real-time receiving and processing around the vehicle state data,Assist early warning algorithm in early warning calculation.The left-turn assistant collision warning algorithm based on V2 X technology can effectively assist the driving of vehicles,and the driver can reduce the collision risk of left-turn to the oncoming vehicle through the warning algorithm.This paper mainly studies collision warning in the case of vehicle left turn.Firstly,a left-turn assisted collision warning algorithm based on V2 X communication is proposed.The algorithm can predict the future trajectory of vehicles according to the state information of other vehicles transmitted by V2 X and map information.Then,the turning intention of the vehicle is judged according to the change of the left and right turn signal and steering wheel Angle of the vehicle.The intersection exit direction is determined according to the map information combined with the turn signal,and the trajectory curve is predicted in real time combining the exit direction and the current state of the vehicle.The cartesian coordinate system is established and the intersection point,namely the collision point,is calculated according to the trajectory equations of the main vehicle and other vehicles.The vehicle is simplified into a rectangular model to calculate whether a collision will occur when the vehicle reaches the collision point.If there is a collision,calculate whether it is within the warning range,and feedback the warning information to the driver if the warning time is appropriate.In order to test the feasibility of the algorithm,the scenario library scheme was first designed and constructed,and the scene was set up through VTD,and the anticollision warning system was jointly constructed through VTD simulation software third-party interface and MATLAB.Then,according to the test requirements,select key scenarios for testing,and judge whether the early warning is based on the received vehicle status information.The experiment tests the influence of road type,lane number,other vehicle behavior and speed on left-turn assisted collision warning algorithm.The results show that the warning accuracy of the algorithm is above 85%in low-speed state,and the accuracy of the correct warning timing is above 80%.The warning accuracy and timing accuracy of 90-degree intersections are above 90%,which is better than that of the intersections with special angles.The number of lanes has little influence on the warning effect.His warning accuracy was higher when the car was going straight than when it was turning.At high speed,the accuracy of vehicle warning algorithm is relatively low,which is about 68.8%.Turn left the study of vehicle effectively through collision risk for early warning of urban intersection,through the formula to calculate the collision risk compared with machine learning reduces the computing burden to shorten the calculation time,by simulating the real trajectory,forecasting curve to calculate collision point,improve the accuracy of the collision point prediction and early warning time,effectively provide early warning information for the driver.