【摘要】 快速扩展随机树算法(rapidly-exploring random trees, RRT)规划移动机器人路径时，存在搜索盲目性强、搜索时间长、收敛速度慢、路径冗余点多且不平滑等问题。鉴于此，提出一种改进的RRT路径规划算法。首先，针对传统RRT算法盲目搜索以及局部极值的问题，提出概率目标偏置与人工势场结合的采样策略，引导随机树的扩展；其次，针对随机树扩展的避障能力差的问题，提出基于安全距离的碰撞检测以及动态变步长扩展策略；最后，针对路径上冗余点多以及曲率不连续的问题，提出考虑安全距离的剪枝优化和三次B样条曲线对初始路径进行拟合优化。仿真结果表明，在不同地图的路径规划中，相比于传统RRT算法，增强了通过狭窄通道能力，优化了路径的平滑性，搜索时间、迭代次数、路径长度分别减少约70%、40%、15%;相比于RRT衍生算法RRT-Connect,搜索时间、路径长度分别减少约25%、10%。更多还原

【Abstract】 When rapidly-exploring random trees algorithm(RRT) plans the path of a mobile robot, has problems such as strong search blindness, long search time, slow convergence rate, multiple redundancy points and uneven path.In order to solve the problems, an improved RRT path planning algorithm is proposed.Firstly, a sampling strategy combining probabilistic target bias and artificial potential field is introduced to guide the expansion of random tree.Secondly, for random tree expansion with poor barrier avoidance, collision detection based on safe distance and dynamic variable step expansion strategy are proposed.Finally, aiming at the problem of many redundant points in the path and discontinuity of curvature, the initial path is fitting optimized with shear optimization and cubic B-spline curve considering safe distance.Simulation results show that in the path planning of different maps, Compared to traditional RRT algorithm, the ability to pass through narrow channels is enhanced.The smoothness of the path is optimized.The search time, iteration times, and path length are reduced by about 70%,40%,15%,respectively.Compared to the RRT-Connect, search time and path length are reduced by about 25% and 10%,respectively.更多还原