基于RRT和DWA的欠驱动UUV路径规划

Underactuated UUV Path Planning Based on RRT and DWA

摘要:

针对无人潜航器(UUV)在未知水下复杂环境的路径规划问题，设计了随机树以及动态窗口的融合算法。该算法基于快速扩展随机树(RRT)以及动态窗口(DWA)两层规划设计，第一层利用随机树算法快速规划出全局路径，在此基础上第二层加载全局路径，针对 UUV 模型的欠驱动和非线性，利用动态窗口算法完成局部路径规划，保证约束条件下 UUV 路径的安全性。通过融合参数 μ 修正内外框架的融合度，有效地弥补了全局路径算法的无法躲避动态障碍物的缺点以及局部路径算法全局能力低下的问题。最后，通过对比仿真验证了融合算法相比于随机树全局算法和动态窗口局部算法的优越性。

Abstract:

Aiming at the path planning of unmanned underwater vehicle(UUV)in unknown underwater complex environment，an RRT-DWA fusion algorithm was designed. The algorithm is based on two-layers planning design of rapid expansion random tree(RRT)and dynamic window(DWA). The first layer uses the RRT algorithm to quickly plan the global path. Based on this and considering the underactuation and non-linearity of the UUV model，the second layer loads the global path，and the DWA algorithm is used to complete the local path planning to ensure the safety of the UUV path under the constraints. The fusion parameters of the inner and outer frames are corrected by the fusion parameters μ，which effectively makes up for the shortcomings of the global path algorithm that cannot avoid dynamic obstacles and improves the low efficiency of the local path algorithm. Finally，the superiority of the RRT-DWA fusion algorithm compared with the RRT global algorithm and the DWA local algorithm is verified through comparative simulation.