含非严格重复扰动的抗扰迭代学习控制  

Anti-disturbance iterative learning control for nonlinear systems with time-iteration-varying disturbances

作　　者：吕庆[1,2] 方勇纯[1,2] 任逍 

机构地区：[1]南开大学机器人与信息自动化研究所 [2]天津市智能机器人技术重点实验室,天津300071

出　　处：《控制理论与应用》2014年第9期1190-1197,共8页Control Theory & Applications

基　　金：国家自然科学基金科学仪器基础研究专款资助项目(61127006)

摘　　要：针对一类含不确定参数及未知扰动的高阶非线性系统,采用类Lyapunov方法,结合部分限幅学习律和滑模控制的优点,提出一种新的滑模鲁棒迭代学习控制算法.根据系统中不确定量的特性,将系统中的不确定性划分为两类:仅沿时间轴变化的不确定性和仅沿迭代轴变化的不确定性.前者采用迭代辨识方法处理,后者采用迭代滑模技术解决.在整个作业区间上,随着迭代次数的增加,控制算法确保系统的跟踪误差收敛到一个界内,控制器信号无抖颤,且闭环系统中其余变量一致有界.当系统扰动仅沿时间轴变化时,系统跟踪误差及其各阶导数沿迭代轴渐近收敛到0,实现系统各个状态的精确跟踪.相比利用连续函数近似法的传统滑模控制,该算法对未知扰动具有更好的鲁棒性.理论证明和仿真结果都说明了该算法的有效性.For a class of higher-order nonlinear systems with parametric uncertainties and unknown disturbances,we propose a novel sliding-mode robust iterative learning control algorithm based on the Lyapunov-like method,which successfully combines the advantages of partially-saturated learning mechanism and sliding mode technique.Uncertainties within the system are classified into two categories,the only time-varying uncertainty and the only iteration-varying uncer tainty.The former is treated by using the iterative identification technique,while the latter is dealt with by employing an iterative sliding mode law.In the entire time interval,it is guaranteed that,along the iteration axis,the designed chatteringfree controller ensures that the tracking error converges to a given bound,while all the remaining signals are bounded.In addition,tracking errors and their derivatives converge asymptotically to zero along the iterative axis in the case of only time-dependent perturbation,which implies accurate tracking for the system states.Compared with the saturationapproximation-based conventional sliding-mode mechanism,the proposed novel control technique presents better robustness against unknown disturbances.Theoretical analysis and simulation results show the effectiveness of the proposed algorithm.

关 键 词：迭代学习控制 滑模控制 参数辨识 收敛性 鲁棒性 抗扰性 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]