基于命令滤波技术的水下机器人位置跟踪控制  

Adaptive Neural Networks Command Filtered Control for AUVs with Input Saturation via Backstepping

作　　者：胡亚强 于金鹏[1] 赵林 于海生[1] HU Yaqiang;YU Jinpeng;ZHAO Lin;YU Haisheng(School of Automation,Qingdao University,Qingdao 266071,China)

机构地区：[1]青岛大学自动化学院,山东青岛266071

出　　处：《青岛大学学报（工程技术版）》2019年第1期95-100,共6页Journal of Qingdao University(Engineering & Technology Edition)

基　　金：国家自然科学基金资助项目(61573204;61573203;61501276;61603204);泰山学者工程专项经费资助(TSQN20161026)

摘　　要：为解决考虑输入饱和的自治水下机器人的位置跟踪问题,本文采用自适应神经网络命令滤波控制方法,设计了对自治水下机器人进行位置跟踪控制器。利用反步法构造考虑输入饱和的控制器,使用神经网络系统逼近自治水下机器人系统中的非线性项,同时采用命令滤波技术解决计算爆炸问题,引入误差补偿机制,降低滤波误差带来的影响。通过选取合适的李雅普诺夫函数求出自适应律,证明闭环系统的所有信号有界。仿真结果表明,该控制器可以使水下机器人系统中的位置跟踪误差收敛到一个足够小的邻域内。该研究具有一定的实际应用价值。In this paper, to solve the position tracking problem in autonomous underwater vehicle (AUV) systems with input saturation, a command filtering based adaptive neural networks backstepping method is proposed. Firstly, the neural networks (NNs) are used to approximate unknown nonlinear function of AUV systems and the backstepping method is employed to design controllers with input saturation. Then, the command filtered control (CFC) method is utilized to deal with the “explosion of complexity” problem, and the filtering error can be reduced properly by the introduction of error compensation mechanism. Finally, by choosing the appropriate Lyapunov function, the adaptive law is obtained, and all the signals of the closed-loop system are proved to be bounded. The simulation results show that the constructed controller can make tracking errors converge to a small region of the origin. This research has some practical application value.

关 键 词：神经网络 命令滤波器 输入饱和 自治水下机器人 

分 类 号：TP242.2[自动化与计算机技术—检测技术与自动化装置]