基于改进快速扩展随机树的机械臂路径规划

针对机械臂路径规划问题,提出一种基于改进RRT算法的路径规划方法。改进RRT结合了目标偏置策略和贪婪生长策略的优点,在随机采样时,以一定概率使采样点偏置为目标节点,降低随机采样的盲目性,在目标节点方向上采用贪婪式扩展策略,增加随机树局部方向上的生长速度。RRT法规划路径结果并非最优,提出改进GPP法删除多余路径节点,优化机械臂运动路径。通过与Biased-RRT和Greedy-RRT数值仿真结果对比,证明了改进RRT在计算时间、迭代次数、扩展节点数上均优于以上方法。在机械臂两种典型工作环境中的仿真结果表明,使用该方法可以较好解决排爆机械臂避障路径规划问题。

Path Planning of Manipulators Based on Improved Rapidly-Exploring Random Tree

A certain improved rapidly-exploring random tree is proposed to solving the path planning problem of EOD manipulators. The improved RRT combines advantages of biased strategy and greedy strategy. By selecting goal node with some probability when sampling,improved RRT reduces the blindness of RRT. Greedy strategy is used for extending trees in the direction of goal node to enhance the extending rapid. The results by RRT are not optimal,an improved GPP technique is developed to prune the redundant nodes of planned paths,optimize the motion path of manipulators. By comparing the numerical simulation results of Biased-RRT and Greedy-RRT,the high efficiency of improved RRT is proved in computation time,iterations and number of extended nodes. The simulation results in two typical environments show that the algorithm can achieved path planning tasks well in obstacle circumstances.