基于数据驱动的范数最优迭代学习控制

Research on Norm Optimal Iterative Learning Control Based on Data Driven

在系统模型确定的前提下,传统的范数最优迭代学习控制(NOILC)可以有效提高伺服系统的跟踪精度。但是在实际控制过程中,系统模型参数往往是变化的,从而导致控制器性能的下降。为此,提出了一种基于数据驱动的范数最优迭代学习控制方法。以系统的输入输出为依据,建立系统估计模型的代价函数,对代价函数进行偏微分处理,得到一种基于数据驱动的非参数模型辨识方法,最后将此模型辨识方法和NOILC相结合。实验结果表明:针对时变系统,此控制方法的跟踪误差为NOILC(Norm optimal iterative learning control,NOILC)的57.1%,并且相比NOILC提前5次收敛。因此,提出的方法能有效改善时变系统的跟踪性能。

The traditional norm optimal iterative learning control(NOILC)can effectively improve the tracking accuracy of the servo system under the premise of determining the system model.However,in the actual control process,the system model parameters are often changed,resulting in a decline in the performance of the controller.Therefore,a data-driven norm-optimal iterative learning control method is proposed.Firstly,based on the input and output of the system,the cost function of the system estimation model is established.Then,the cost function is processed by partial differential,and a data-driven nonparametric model identification method is obtained.Finally,the model identification method is combined with NOILC.The experimental results show that for the time-varying system,the tracking error of this control method is 57.1%of NOILC,and it converges five times ahead of NOILC.Therefore,the proposed method can effectively improve the tracking performance of time-varying systems.