为实现移动机器人在复杂动态障碍物环境中的避障,提出一种改进的快速随机扩展树(rapidly-exploring random tree,RRT*)与动态窗口法(dynamic window approach,DWA)相融合的动态路径规划方法.基于已知环境信息,利用改进RRT*算法生成全局最优安全路径.通过消除RRT*算法产生的危险节点,来确保全局路径的安全性;使用贪婪算法去除路径中的冗余节点,以缩短全局路径的长度.利用DWA算法跟踪改进RRT*算法规划的最优路径.当全局路径上出现静态障碍物时,通过二次调整DWA算法评价函数的权重来避开障碍物并及时回归原路线;当环境中出现移动障碍物时,通过提前检测危险距离并转向加速的方式安全驶离该区域.仿真结果表明:该算法在复杂动态环境中运行时间短、路径成本小,与障碍物始终保持安全距离,确保在安全避开动态障碍物的同时,跟踪最优路径.
Dynamic Path Planning for Mobile Robot Based on RRT* and Dynamic Window Approach
A dynamic path planning method combining RRT* and dynamic window approach(DWA)is proposed to realize the obstacle avoidance of mobile robot in complex environment of dynamic obstacles.Improved RRT* algorithm is used to generate the global optimal safe path based on the known environment information.By eliminating the dangerous nodes generated by RRT* algorithm,the security of global path is ensured.Greedy algorithm is used to remove the redundant nodes in the path to reduce the length of global path.DWA is used to track along the global optimal path planned by the improved RRT* algorithm.When static obstacles appear on the global path,the weights of DWA algorithm evaluation function is adjusted twice to avoid the obstacles and return to the original path in time.When moving obstacles is in the environment.By detecting the dangerous distance in advance,changing direction and speeding up,the robot can safely drive away from the area.Simulation experiments verify that the improved fusion algorithm proposed in complex dynamic environment has shorter running time,smaller path cost,and always keeps safe distance from obstacles,which ensures the optimal tracked path while safely avoiding the dynamic obstacles.
mobile robotpath planningimproved RRT* algorithmdynamic window approachdynamic obstacle avoidance
NETL
NSTL
万方数据
