基于RRT改进算法的AGV路径规划

AGV Path Planning Based on Improved RRT Algorithm

传统快速扩展随机树(RRT)算法在搜索空间中,随机采样生成我们所需要的树,由树的起始点直到终点,探索出一条无障碍的路径。采样点是均匀随机,导致算法过于随机,生成路径的效率不高且生成路径质量偏低,在面对狭窄通道时容易导致算法局部循环甚至搜索失败,传统算法生成的路径过于曲折不利于跟踪行驶。针对这些问题,改进后的算法在RRT的基础上,增加算法贪婪计算和目标节点的启发;将扩展的采样点重点集中于一定的区域,满足正态分布。仿真实验表明,改进后的算法效率更高,生成路径质量高,面对狭窄通道这个传统难题也可以高质高效地生成一条路径,利于AGV跟踪行驶。

The traditional fast expanding random tree(RRT)algorithm generates the tree we need by random sampling in the search space,and explores an obstacle free path from the start point to the end point of the tree.The sampling points are uniform and random,which leads to the algorithm is too random,the efficiency of generating path is not high and the quality of generated path is low.When facing the narrow channel,it is easy to cause the algorithm local cycle or even search failure.The path generated by tradi⁃tional algorithm is too tortuous,which is not conducive to tracking.In order to solve these problems,based on RRT,the improved algorithm adds greedy computation and the heuristic of target nodes,and focuses the extended sampling points on a certain area to meet the normal distribution.Simulation results show that the improved algorithm has higher efficiency and higher quality of generat⁃ed path.In the face of the traditional problem of narrow channel,it can also generate a path with high quality and efficiency,which is conducive to AGV tracking.