文献类型：期刊文献

英文题名：Distributed cooperative H_∞ optimal control of underactuated autonomous underwater vehicles based on reinforcement learning and prescribed performance

作者：Zhuo, Jiaoyang[1,2];Tian, Xuehong[1,2,3];Liu, Haitao[1,2,3]

机构：[1]Guangdong Ocean Univ, Sch Mech Engn, Zhanjiang 524088, Peoples R China;[2]Guangdong Ocean Univ, Shenzhen Inst, Shenzhen 518120, Peoples R China;[3]Guangdong Engn Technol Res Ctr Ocean Equipment & M, Zhanjiang 524088, Peoples R China

年份：2024

卷号：312

外文期刊名：OCEAN ENGINEERING

收录：SCI-EXPANDED(收录号：WOS:001328773600001)、、EI(收录号：20243917107388)、Scopus(收录号：2-s2.0-85204762773)、WOS

基金：This work was supported by the Guangdong Basic and Applied Basic Research Foundation [grant number 2024A1515011345] , the Key Project of the Department of Education of Guangdong Province [2023ZDZX1005] , the Shenzhen Science and Technology Program [grant number JCYJ20220530162014033] , the National Natural Sci-ence Foundation of China [grant number 62171143] , and the Science and Technology Planning Project of Zhanjiang City [grant numbers 2021A05023 and 2021E05012] .

语种：英文

外文关键词：Underactuated autonomous underwater vehicle; Optimal control; Trajectory tracking; Prescribed performance control; Reinforcement learning; H-infinity control

外文摘要：To balance energy resources and control performance, an H-infinity optimal control method based on prescribed performance control (PPC) and a reinforcement learning (RL) algorithm with actor-critic mechanisms for distributed cooperative control is proposed for multiple five-degree-of-freedom underactuated autonomous underwater vehicles (AUVs) with unknown uncertainty disturbances. First, an optimal control strategy combining PPC is proposed to achieve optimal control of a cooperative system while ensuring that the error always stays within the prescribed boundary. Second, to suppress uncertainty disturbances, H-infinity control methods are proposed to improve the robustness of the system. Achieving H-infinity optimal control requires solving the Hamilton-Jacobi-Bellman (HJB) equation, but the inherent nonlinearity of the HJB equation makes it difficult to solve. Therefore, an adaptive approximation strategy incorporating an online RL method with an actor-critic architecture is used to solve the above problem, which dynamically adjusts the control strategy to ensure system control performance through the environment assessment-feedback approach. In addition, a distributed adaptive state observer is proposed to obtain information about the virtual leader for each agent so that leader information can be accurately obtained, even if the agent communicates only with neighboring agents. Using the above control method, all errors of the formation system are proven to be uniform and ultimately bounded according to Lyapunov's stability theorem. Finally, a numerical simulation is performed to further demonstrate the effectiveness and feasibility of the proposed method.