The Parallel Driving System Based On Improved Multiple Vehicles Collision Avoidance Algorithms

In recent years,with the rapid development of the artificial intelligence and the computer technology and the people's yearning for a better life,the intelligent industry has received extensive attention.As an indispensable means of transportations for people to travel,the vehicles have also been pushed into the intelligent industry.Autonomous vehicles can solve fuel problems to a great extent and avoid dangerous accidents.However,there are still some difficulties in people's life that need to be overcome.For example,the influence of weather on sensor equipment leads to abnormal driving of the vehicle,so the parallel driving system fully solves this problem in the excessive process from the driving assistance system to the fully automatic driving.The parallel driving system consists of automatic driving vehicle and the cloud-control system.The cloud-control system will be takeover the autonomous vehicles when the vehicle in emergency,and the trajectory planning algorithm of the virtual model in the virtual scene will be used to re-plan and control the real vehicle,so as to ensure the safe driving of the autonomous vehicle in the emergency.The reason why the trajectory planning algorithm is introduced into the concept of parallel driving is that the parallel driving system includes not only real automatic driving vehicles,but also the operation of virtual simulation scenes.The two can correct each other,which improves the fault tolerance rate of the system and ensures the safe and stable driving of the automatic driving system.In this paper,the related research is studied firstly based on the multiple vehicles collision avoidance algorithms(MVCA).The algorithm firstly performs the global path planning based on the A* algorithm,and obtains the global optimal path of the automatic driving as the vehicle reference trajectory,and then based on the algorithm of Optimal Reciprocal n-body Collision Avoidance(ORCA)performs local path planning,which is mainly used to input the surrounding obstacles acquired by sensor devices such as lidar and millimeter wave radar as the algorithm input during the driving of the autonomous vehicle,and then the real-time trajectory of the vehicle is updated to avoid collision and complete the automatic driving of the vehicle.Then the paper first made an offline test according to the algorithm,and then after the offline test results meet the requirements,the algorithm was tested in real vehicle.In the real vehicle test,the communication between the devices passed by the controller area network(CAN).After many experiments,it can be confirmed that the vehicle will drive automatically through the algorithm,and the trajectory is the same as the expected trajectory.Finally,we first introduce the virtual scene modeling and virtual interaction unit in the parallel driving system.Then we complete the parallel driving system based on the virtual scene and automatic driving,including the cloud control system.In this system,we not only do video transmission through the 4G network,but also do the test by 5G network in the strategic cooperation between the company and China Mobile.The test shows that the communication effect is good on the basis of 5G network.