基于改进RRT*算法的自动驾驶车辆路径规划研究

Path planning of autonomous vehicles at intersections based on improved RRT*

针对自动驾驶车辆,文章在交叉路口环境下提出了一种改进的快速搜索随机树(RRT*)路径规划算法。首先,对自动驾驶车辆的驾驶行为环境予以描述;其次,针对原始RRT*算法提出改进的目标偏向策略予以改善;进一步,对原始RRT*算法在交叉路口无效采样的问题,提出一种概率采样策略。基于Matlab/Simulink联合仿真平台构建相应环境使进行车辆直行驾驶,所规划路径长度为100.35m,仿真时长为5.71s。

For autonomous vehicles,this paper proposes an improved Rapidly-exploring Random Tree star(RRT*)path planning algorithm in the intersection environment.First of all,this paper describes the driving environment of autonomous vehicles.Secondly,the improved target bias strategy is adopted Furthermore,for the problem of invalid sampling of the original RRT*algorithm at intersections,this paper proposes a sampling strategy based on the expected generation probability of sampling points.Based on the MATLAB/Simulink joint simulation platform,the corresponding environment is constructed to make the vehicle drive straight.The length of the planned path is 100.35m and the simulation duration is 5.71s.