基于输入饱和抑制和神经网络的欠驱动AUV轨迹跟踪滑模控制  

Anti-windup and Neural Networks-Based Sliding Mode Control of Trajectory Tracking for Underactuated AUV

作　　者：刘帅 罗伟林 LIU Shuai;LUO Weilin(Research and Development Center of Marine Engineering Equipment,Fuzhou Institute of Oceanography,Fuzhou 350108,China;College of Mechanical Engineering and Automation,Fuzhou University,Fuzhou 350108,China)

机构地区：[1]福州海洋研究院海洋工程装备设计制造研发中心,福州350108 [2]福州大学机械工程及自动化学院,福州350108

出　　处：《中国造船》2023年第6期240-249,共10页Shipbuilding of China

基　　金：福州海洋研究院科技项目(2022F13)。

摘　　要：提出了一种基于输入饱和抑制和神经网络的欠驱动AUV非奇异快速积分终端滑模控制器。用视线制导法定义期望跟踪速度,针对速度跟踪误差用非奇异快速积分终端滑模控制方法设计动力学控制器。用自适应RBF神经网络逼近时变干扰和计算难以获得的期望速度微分项,设计了抗饱和控制器解决执行器饱和问题。通过数值仿真,验证了提出的控制器对外界干扰具有鲁棒性,并能较好地处理输入饱和的影响。An underactuated AUV non-singular fast integrating terminal sliding mode controller with input saturation constraint and external disturbance is proposed.Firstly,the desired longitudinal velocity and expected yaw rate are obtained by the line-of-sight guidance method.Aiming at the velocity tracking errors and angular velocity tracking errors,a trajectory tracking controller is designed by using the non-singular fast integrating terminal sliding mode control method.The adaptive RBF neural network is used to approximate unknown time-varying disturbances and the time derivatives of desired velocity.A kind of anti-windup controller is designed to solve the actuator saturation problem.Numerical simulation shows that the proposed controller is robust to external disturbance and can handle the input saturation effect well.

关 键 词：欠驱动自主水下机器人 自适应RBF神经网络 非奇异快速终端滑模控制 轨迹跟踪控制 输入饱和 

分 类 号：U664.82[交通运输工程—船舶及航道工程]