基于BP算法的无模型自适应迭代学习控制  

Model-free adaptive iterative learning control based on BP algorithm

作　　者：李佳[1] 

机构地区：[1]青岛科技大学自动化与电子工程学院自主导航与智能控制研究所,山东青岛266042

出　　处：《微型机与应用》2010年第22期83-85,共3页Microcomputer & Its Applications

基　　金：国家自然科学基金(60974040)

摘　　要：为了改善针对一般非线性离散时间系统的控制性能,引入"拟伪偏导数"概念,给出了一般非线性离散时间系统沿迭代轴的非参数动态线性化形式,并综合BP神经网络以及模糊控制各自的优点,提出了基于BP算法无模型自适应迭代学习控制方案。仿真结果表明,该控制器对模型有较强的鲁棒性和跟踪性。In order to improve the system for the general nonlinear discrete-time control performance,a non-parameter dynamic linearization along the iteration axis of the general nonlinear discrete-time systems is developed by introducing the concept of ＂Mimic Pseudo Partial Derivative＂,BP algorithm is proposed based on Model-free adaptive iterative learning control program integration of BP neural network and fuzzy control of their respective advantages.Simulation results show that the controller model has strong robustness and perfect performance of tracking.

关 键 词：BP算法 无模型自适应控制 迭代学习 拟伪偏导数 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]