Research On Autonomous Vehicle Motion Planning Algorithm Under Road Environment And Parking Scene

In recent years,autonomous driving technology has attracted more and more people’s attention,becoming the focus of the automotive field.The development of autonomous vehicles can alleviate traffic pressure,achieve energy saving,reduce emission,and make mobile travel more convenient.Therefore,major companies and scientific research institutes have invested manpower and material resources to develop autonomous driving technology.The development of autonomous driving technology is quite difficult,because it contains too many key technologies,such as environmental perception,decision planning,motion control,etc.,each of the key technologies has many problems that need to be solved urgently.Motion planning is one of the key technologies for autonomous driving,and how to generate a collision-free,time-saving,comfortable,reliable,robust,and adaptable trajectory is crucial.Among them,the road environment and the parking scene are the two most common environments for vehicles to travel.Therefore,this paper studies the design of the motion planning of autonomous vehicles against these two backgrounds.First,the overall framework and decision-making framework of autonomous vehicles are selected,and the design criteria and constraints of the motion planning are analyzed.Second,based on the road environment,the autonomous vehicle motion planning algorithm is studied.Due to the large number of traffic participants in the road environment and the high speed of vehicles,the real-time requirements of the motion planning algorithm are relatively high,so this paper uses a path-speed decoupled trajectory planning framework.First,the local target points are laterally sampled,and the path cluster is generated using the cubic spiral curve as the path model;Secondly the velocity and acceleration are sampled on each path,and the speed cluster is generated with the cubic polynomial curve;Finally,according to the defined static and dynamic costs,each trajectory is evaluated,and the trajectory with the lowest total cost is selected as the trajectory executed by the lower-level control tracking module.And through the typical road scene to verify the developed algorithm,the simulation results show that the road planning algorithm developed in this paper can well complete the common driving tasks in the road environment.Third,based on the parking scene,the autonomous vehicle motion planning algorithm is studied.In this paper,two automatic parking motion planning algorithms are developed: the first is based on the path-speed decoupling method,the path information is obtained by the hybrid A~* algorithm,and the speed planning is based on the convex optimization method on this path;the second is based on the method of optimal control,the parking problem is modeled as an optimal control problem and the first method is used to obtain the result as the initial solution to speed up the solution.The feasibility of the algorithm is verified by parallel parking and vertical parking,and the second method is superior to the first method in drivability.Finally,I build an unmanned electric wheelchair platform independently developed by the laboratory,and set up experimental scenes on the campus of Chongqing University to verify the feasibility of the motion planning algorithm under the road environment.