多AUV的自组织人工势场编队控制方法研究  

Research on the Method of Multi-AUV Formation Control Based on Self-organized Artificial Potential Filed

作　　者：陈杨杨 朱大奇[1] 李欣[1] CHEN Yang-yang;ZHU Da-qi;LI Xin(Shanghai Engineering Research Center for Intelligent Maritime Search and Rescue and Underwater Vehicle, Shanghai MaritimeUniversity, 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

出　　处：《控制工程》2019年第10期1875-1881,共7页Control Engineering of China

基　　金：国家自然科学基金课题(U1706224,91748117,51575336);上海市科委创新行动计划(18JC1413000,18DZ2253100,16550720200)

摘　　要：针对多自治水下机器人(Autonomous Underwater Vehicles, AUV)水下编队与安全避障问题,将自组织映射(Self-Organizing Map, SOM)神经网络和人工势场方法相结合,提出多AUV自组织人工势场编队控制方法。首先,根据领航AUV的位置产生虚拟AUV位置,将虚拟AUV位置作为SOM的输入向量进行竞争计算,输出为跟随AUV的位置,从而控制跟随AUV到达期望目标点;接着考虑编队行进中的避障问题,采用人工势场法进行避障,规划编队路径。最后通过仿真给出算法的有效性说明。In this paper, self-organizing map(SOM) and artificial potential field methods are combined to solve the multi-AUV(autonomous underwater vehicle) formation and obstacle avoidance problem. An self-organized artificial potential filed formation control method is then proposed. First of all, according to the location of the leader AUV, virtual AUVs’ positions are generated. Virtual AUV positions are used as the input vectors of the SOM network for calculation. The output are the positions of the follower AUVs which are able to control the follower AUVs to reach the desired target points. Secondly, considering the formation problem of obstacle avoidance, the artificial potential field method is used to avoid obstacle and re-plan paths for formation. Finally, the effectiveness of the proposed algorithm is verified by simulations.

关 键 词：多水下机器人 自组织映射神经网络 编队控制 人工势场 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]