一种新型多AUV水下围捕路径规划算法  

A Novel Path Planning Algorithm for Multi-AUV Underwater Hunting

作　　者：蒋骁迪 甘文洋 JIANG Xiao-di;Gan Wen-yang(Shanghai Engineering Research Center of Intelligent Maritime Search&Rescue and Underwater Vehicles,Shanghai Maritime University,Shanghai 201306,China)

机构地区：[1]上海海事大学智能海事搜救与水下机器人上海工程技术研究中心,上海201306

出　　处：《计算机仿真》2021年第9期376-380,428,共6页Computer Simulation

基　　金：上海市科技创新行动计划(18JC1413000,18DZ2253100);国家自然科学基金重大研究计划(91748117)。

摘　　要：针对二维水下环境下的多水下机器人(Autonomous Underwater Vehicle,AUV)协作围捕问题,提出一种基于改进生物启发神经网络和位置分配的围捕策略。首先对二维水下环境进行栅格地图的构建;然后多个围捕AUV之间相互合作,提出利用"位置分配"策略合理分配逃逸目标AUV周围的围捕占位点给围捕AUV;最后利用改进的Glasius生物启发神经网络算法来完成围捕AUV到达围捕点的路径规划任务,捕获逃逸目标AUV,完成协作围捕任务。与传统生物启发神经网络动态分配围捕策略相比,二维水下静态障碍物环境下的仿真验证了上述协作围捕算法能够在水下障碍物的工作空间中高效完成围捕任务。In order to solve the problem of multi-AUVs in a two-dimensional underwater environment,a hunting strategy based on an improved bio-inspired neural network and location assignment is proposed.Firstly,a grid map was constructed for a two-dimensional underwater environment.Then,it was proposed to use the strategy of"location-allocation"to reasonably allocate the capture sites around the escape target AUV to the hunting AUVs.And finally,the improved Glasius bio-inspired neural network algorithm was used to complete the path planning task of hunting AUVs to the capture sites.The escape target AUV was captured to complete the cooperative hunting task.Compared with the traditional bio-inspired neural network dynamic allocation hunting strategy,the simulation experiment under the two-dimensional underwater static obstacle environment proves that the proposed cooperative hunting algorithm is capable of efficiently completing the hunting task in the working space of the underwater obstacle.

关 键 词：多水下机器人协作围捕 围捕占位 路径规划 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]