摘要
本文针对现有的一类小型观测级ROV(Remotely Operated Vehicle)的四自由度运动,提出一种基于径向基函数神经网络(Radial Basis Function Neural Network,RBFNN)自适应滑模控制方法。考虑了ROV运动时模型不确定因素的干扰,并给出ROV的模型不确定项。通过引入RBF神经网络算法,对ROV模型不确定项进行了补偿。此外,采用反正切函数替换了经典滑模控制中常用的符号函数,减小了滑模控制的抖动。通过Lyapunov稳定性定理,验证了该系统是全局渐进稳定的。使用Matlab/Simulink仿真软件,证明了RBF神经网络自适应滑模控制方法的可行性。
This work focus on a RBFNN(Radial Basis Function Neural Network)based adaptive sliding mode control(RBFSMC)for a kind of small monitoring ROV in 4 degrees of freedom motion.In this paper,the ROV model uncertainties were considered,and the mathematical model of the ROV model uncertainties were given.By using RBF neural network,the ROV model uncertainties were compensated.In addition,for reducing the chattering of conventional sliding mode control(CSMC),the sign function in CSMC was replaced by arctan function.According to Lyapunov stability theorem,the global asymptotic stability of the system was proven.The Matalb/Simulink experiment results validate the effectiveness of RBFSMC.
作者
杨淼
盛智彬
王海文
殷歌
YANG Miao;SHENG Zhi-bin;WANG Hai-wen;YIN Ge(The Department of Electronic Engineering of Jiangsu Ocean University,Lianyungang 222002,China;The Department of Mechanical and Ocean Engineering of Jiangsu Ocean University,Lianyungang 222002,China;Institute of Marine equipment,Jiangsu University of Science and Technology,Zhenjiang 212000,China)
出处
《舰船科学技术》
北大核心
2020年第10期83-89,共7页
Ship Science and Technology
基金
国家自然科学基金青年项目(61601194)
江苏科技大学海洋装备研究院高技术协同创新项目(HZ20190005)
2019年江苏省研究生科研创新项目(KYCX19_2314)。
