基于改进RRT-Connect算法的全局路径规划

Global Path Planning Based on Improved RRT-Connect Algorithm

针对RRT-Connect算法在复杂环境内的路径规划中存在探索性弱、收敛速度慢、冗余节点多、搜索路径较长等问题,提出一种改进的RRT-Connect算法。通过引入高质量随机点和动态步长的方法,提高了生成随机树的质量并减少了冗余节点数量;采用正向寻优和逆向贪婪的方式,改善了搜索路径较长的问题。实验结果表明,改进RRT-Connect算法平均路径规划时间缩短26.41%,平均路径规划长度缩短19.05%,平均路径规划节点个数减少41.91%,证明了改进RRT-Connect算法相比于原算法规划效率更高,规划时间更少,规划路径质量更优。

Aiming at the problems of RRT-Connect algorithm in path planning in complex environment,such as weak exploration,slow convergence,many redundant nodes and long search path,an improved RRT-Connect algorithm is proposed.By introducing high quality random points and dynamic step size method,the quality of generating random tree is improved and the number of redundant nodes is reduced.The method of forward searching and reverse greedy is adopted to solve the problem of long search path.The experimental results show that the average path planning time of the improved RRT-Connect algorithm is shortened by 26.41%,the average path planning length is shortened by 19.05%,and the average number of path planning nodes is reduced by 41.91%,which proves that the improved RRT-Connect algorithm has higher planning efficiency and less planning time than the original algorithm.The quality of the planned path is better.