复杂环境下DWA与RRT算法融合的AUV局部路径规划

Local path planning for AUV with fusion of DWA and RRT algorithms in a complex environment

针对复杂水下环境下的自主水下航行器(autonomous underwater vehicle,AUV)局部路径规划问题,传统动态窗口法(dynamic window approach,DWA)存在复杂障碍物中陷入局部停滞,动态避障性能不佳等问题,本文提出了一种基于DWA与快速随机搜索树(rapid-exploration random tree,RRT)算法融合的路径规划算法。改进的DWA算法速度空间根据整个动态窗口的周期生成,重设了评价函数并结合AUV任务环境引入洋流能耗评价函数;改进的RRT算法在局部已知空间内规划导引点,帮助DWA脱离局部停滞状态并实现更安全的动态避障。将2种算法融合,实现了AUV在复杂水下环境中的局部路径规划。仿真表明,该融合算法能够降低AUV在洋流中的能耗代价,解决了DWA在复杂障碍物中陷入局部停滞的问题,能够安全有效地躲避动态避障物。

For the local path planning problem of autonomous underwater vehicle(AUV)in a complex underwater environment,traditional dynamic window approach(DWA)has the problems of getting into local stagnation in complex obstacles and poor dynamic obstacle avoidance performance,etc.In this paper,we propose a path planning algorithm based on the fusion of DWA and Rapid-exploration random tree(RRT)algorithms.The improved DWA algorithm generates the velocity space based on the whole dynamic window period,resets the evaluation function and introduces the evaluation function of ocean current energy consumption in an AUV mission environment;the improved RRT algorithm plans the guide points in a local known space,which helps DWA to get out of the local stagnation and achieve a safer dynamic obstacle avoidance.The two algorithms are fused to achieve local path planning for AUV in a complex underwater environment.Simulations show that the fusion algorithm can reduce the energy cost of AUV in ocean currents,solve the problem of DWA getting into local stagnation in complex obstacles,and can avoid dynamic obstacles safely and effectively.