小脑神经网络用于不确定时滞系统的鲁棒非脆弱控制  

Application of CMAC to robust non-fragile control of systems with time-delay and uncertainties

作　　者：付兴建[1] 郭宏梅 FU Xing-jian;GUO Hong-mei(School of Automation,Information Science and Technology University,Beijing 100192,China)

机构地区：[1]北京信息科技大学自动化学院,北京100192

出　　处：《西安科技大学学报》2020年第3期477-483,共7页Journal of Xi’an University of Science and Technology

基　　金：国家自然科学基金(61573230);北京信息科技大学促进高校内涵发展科研水平提高项目(5211910950);北京市教委科技计划项目(KM201811232013)。

摘　　要：针对不确定时滞系统的鲁棒跟踪控制问题,设计了一种基于小脑神经网络CMAC的鲁棒非脆弱控制器。首先,给出小脑模型神经网络控制系统的算法。其次针对一类不确定时滞系统,根据李雅普诺夫稳定理论,进行了鲁棒非脆弱控制器的设计。假设反馈控制中即含有状态反馈不确定性,也具有状态时滞的不确定性。证明不确定时滞系统鲁棒非脆弱控制存在的条件。该条件可以利用Matlab的线性矩阵不等式LMI工具箱来求解鲁棒控制器的参数。之后利用CMAC神经网络较强的学习能力和鲁棒非脆弱控制器对参数摄动抑制作用的特点,将鲁棒非脆弱控制器与小脑模型神经网络CMAC相结合,构成小脑模型神经网络与鲁棒非脆弱控制器的复合控制,实现对不确定时滞系统的跟踪控制。仿真结果显示,对于输入端扰动和一定程度的参数摄动,经过复合控制器的作用,被控系统能在短时间的抖动后逐渐趋于稳定,不仅具有较快的响应速度,还具有较短的收敛时间和令人满意的跟踪精度。该种复合控制表现出较强抗干扰能力及鲁棒性。The robust non-fragile controller based on CMAC neural network was designed for the tracking control problem of uncertain time-delay systems.Firstly,the algorithm of the cerebellar model neural network control system was put forward.Secondly,the tracking control problem was transformed into a robust control problem.For a class of uncertain time-delay systems,the robust non-fragile controller was designed according to Lyapunov stability theory.It was assumed that the feedback control contained state feedback uncertainty and uncertainty of state delay feedback.The conditions for the existence of a robust non-fragile controller were proved.It was possible to solve the parameters of the robust controller by using the linear matrix inequality LMI toolbox in Matlab.Then,by using the CMAC neural network with strong learning ability and the characteristics of robust non-fragile controller for parameter perturbation suppression,the two controllers were combined to form the composite control of the cerebellar model neural network and the robust non-fragile controller.Tracking control for uncertain time-delay systems was realized.The simulation results show that for the input disturbance and a certain degree of parameter perturbation,the controlled system can be gradually stabilized after a short period.Through the composite controller,the system has a faster response speed,a shorter convergence time and satisfactory tracking accuracy.This kind of composite control shows strong anti-interference ability and robustness.

关 键 词：小脑模型神经网络 非脆弱控制 不确定性 时滞 线性矩阵不等式 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]