执行器故障和未知控制方向系统的自适应模糊学习控制  

Adaptive Fuzzy Learning Control for Systems with Actuator Failures and Unknown Control Directions

作　　者：许清媛 郑创涛 麦庆全 万凯 何婉滢 XU Qingyuan;ZHENG Chuangtao;MAI Qingquan;WAN Kai;HE Wanying(School of Electronic and Information,Guangdong Polytechnic Normal University,Guangzhou 510665,China;School of Electronic Information and Electrical Engineering,Huizhou University,Huizhou 516007,China)

机构地区：[1]广东技术师范大学电子与信息学院,广东广州510665 [2]惠州学院电子信息与电气工程学院,广东惠州516007

出　　处：《控制工程》2023年第9期1706-1719,共14页Control Engineering of China

基　　金：国家自然科学基金资助项目(62103159);广东省基础与应用基础基金项目(2020A1515110414,2022A1515140019);广东省普通高校重点领域专项项目(2019KZDZX1042);广东省普通高校青年创新人才项目(2021KQNCX093)。

摘　　要：在工业生产中,工业机器人等设备在重复运动中容易出现执行器故障,进而降低生产效率。另外,考虑到很多实际系统无法提前预知控制方向,因此,面向执行器故障和未知控制方向系统,提出了一种自适应模糊迭代学习控制(iterative learning control,ILC)算法。首先,使用模糊逻辑系统(fuzzy logic system,FLS)估计期望的控制信号,并设计额外的自适应项来补偿执行器故障和系统未知函数带来的副作用。其次,考虑到系统的控制方向是未知的,采用离散Nussbaum型函数在迭代方向对其进行辨识,并将该函数应用于自适应模糊ILC算法。最后,通过Lyapunov-like函数证明了所提自适应模糊ILC算法的可行性:当迭代次数趋于无穷大时,除t∈{0,1,…,m-1}时刻之外的ILC跟踪误差可收敛到一个可调界内,且所有系统信号保持有界。仿真结果验证了所提方法的可行性和有效性。In industry production,equipment such as industrial robots are susceptible to actuator failures during repetitive motion,which in turn reduce productivity.In addition,considering that many practical systems cannot know the control directions in advance,an adaptive fuzzy iterative learning control(ILC)algorithm is proposed for systems with actuator failures and unknown control directions.Firstly,a fuzzy logic system(FLS)is used to estimate the desired control signal and an additional adaptive item is designed to compensate for the side effects caused by actuator failures and unknown functions of the controlled system.Secondly,considering that the control direction of the system is unknown,a discrete Nussbaum-type function is used to identify it along the iterative axis and applied to the adaptive fuzzy ILC algorithm.Finally,the feasibility of the proposed adaptive fuzzy ILC algorithm is demonstrated by a Lyapunov-like function.When the number of iterations approaches infinity,the ILC tracking errors converge to an adjustable residual set except for the times t∈{0,1,…,m-1},and all system signals remain bounded.The simulation results verify the feasibility and effectiveness oftheproposed scheme.

关 键 词：自适应模糊迭代学习控制 执行器故障 未知控制方向 离散Nussbaum型函数 模糊逻辑系统 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]