Research On Path Planning Method Of Autonomous Ground Vehicle Based On RRT

In recent years,the rapid development of artificial intelligence technology takes a feasible solution for autonomous driving.The potential value of unmanned driving technology is enormous.It can effectively improve traffic congestion,reduce traffic accidents and environmental pollution.The main research object of this thesis is the path planning of driverless vehicles.The main goal of this thesis is to obtain collision-free paths satisfying the normal trajectory of vehicles quickly.Rapidly-exploring Random Tree(RRT)is a sampling-based path planning algorithm.It has been widely used in the robot field because of its fast planning speed.However,when applied in the field of unmanned driving,there are still some shortcomings,such as the path does not meet the vehicle feasibility requirements,the path is not optimal,and so on.Therefore,it is necessary to further study it.In this thesis,a bidirectional asymptotically optimal Rapidly-exploring Random Tree(B-RRT*)algorithm is proposed to solve the problem of path planning in the field of driverless vehicles.The main advantage of the algorithm is that it can quickly plan the optimal path satisfying various constraints of vehicles.The main work of this thesis is as follows:Firstly,the Rapidly-exploring Random Tree is improved.By analyzing the constraints of vehicles and the problems existing in the application of Rapidly-exploring Random Tree in the field of autonomous driving,the methods of pruning sampling space and enlarging the target node area are proposed to speed up the search and improve the search efficiency;the regularized vehicle processing method is proposed to effectively avoid invalid collision detection;and the pruning method is proposed to remove redundant turning points to make the path satisfies the constraint of maximum steering angle.Using the B-spline curve to smooth the path makes the path satisfy the constraint of continuous curvature.The simulation results show that the improved algorithm satisfies the feasibility requirements of the vehicle.Secondly,aiming at the defect of non-optimal or sub-optimal path planning of improved RRT algorithm,a B-RRT*algorithm is proposed.Bidirectional search is used to speed up the search and improve the search efficiency;asymptotic optimality is used to overcome the blindness of random sampling.The heuristic search is improved to adapt to the bidirectional search mechanism of the algorithm,by analyzing the parameters that affect the algorithm,the range of parameters that make the planning time shorter and the path quality better is obtained.Finally,the proposed algorithm is simulated and analyzed.Under different environment models,the algorithm is simulated and compared by using MATLAB software.The superiority and validity of B-RRT*algorithm is analyzed by analyzing the two indicators of planning time and path quality.The simulation results show that the algorithm can quickly plan the asymptotically optimal path satisfying various constraints of vehicles,and effectively solve the problem of path planning for autonomous ground vehicles.