海洋自主水面船舶跨水域自适应神经控制  

Adaptive neural control for marine autonomous surface ships in cross-water scenarios

作　　者：叶翔 陈超 贾建雄[2] 陈航 YE Xiang;CHEN Chao;JIA Jian Xiong;CHEN Hang(School of Naval Architecture and Maritime,Zhejiang Ocean University,Zhoushan 316022,China;Zhejiang Branch of China Classification Society,Ningbo 430060,China)

机构地区：[1]浙江海洋大学船舶与海运学院,浙江舟山316022 [2]中国船级社浙江分社,浙江宁波430060

出　　处：《中国舰船研究》2025年第1期309-316,共8页Chinese Journal of Ship Research

基　　金：浙江省“尖兵”“领雁”研发攻关计划资助项目(2023C03181);浙江海洋大学企业行业难题攻关资助项目(1118106412301,1118106412204)。

摘　　要：[目的]针对跨水域场景下海洋自主水面船舶受模型参数不确定和外界环境干扰未知的跟踪控制问题,提出一种具有指定性能的自适应神经控制方案。[方法]在反步法设计框架下,利用神经网络逼近模型参数不确定和未知的外界环境扰动,构造一种新的指定性能函数,并结合障碍李雅普诺夫函数来实现跨水域设计的转换,同时使用动态面控制技术降低系统计算的复杂度,借助李雅普诺夫理论进行稳定性分析,证明控制系统内所有信号都是有界的。[结果]仿真结果表明,所提控制方案能够解决海洋自主水面船舶跨水域跟踪控制,且跟踪误差能够满足在离线预定义时间内收敛至给定的约束范围。[结论]所做研究能够解决船舶的跨水域跟踪控制问题,为受限水域船舶的跟踪控制提供参考价值,且具有实际的工程意义。[Objective]An adaptive neural control(ANC)scheme with specified performance is proposed for the tracking control of marine autonomous surface ships(MASS)subject to uncertain model parameters and unknown external environmental disturbances in cross-water scenarios.[Methods]Under the backstepping design framework,a neural network is utilized to approximate the uncertain model parameters and unknown external environmental disturbances.A novel specified performance function is constructed and combined with the barrier Lyapunov function(BLF)to transform the cross-water design,while the dynamic surface control technique is employed to reduce the system's computational complexity.Stability analysis is then performed by means of Lyapunov theory to demonstrate that all signals within the control system are bounded.[Results]The simulation results show that the designed control scheme is not only capable of solving the cross-water tracking control of MASS,but that the tracking error can satisfy the convergence to a given bounded range within a predefined time offline.[Conclusion]The results of this study can solve the cross-water tracking control problems of MASS and provide valuable references for the tracking control of ships in restricted waters,giving them practical engineering significance.

关 键 词：无人船 海洋自主水面船舶 神经网络 自适应神经控制 障碍李雅普诺夫函数 跨水域场景 

分 类 号：U675.91[交通运输工程—船舶及航道工程] U664.82[交通运输工程—船舶与海洋工程]