Active iterative learning control based on big data

In this paper, we introduce the concept of data similarity for a class of nonlinear time-varying discrete time systems. By combining the large data processing method with the iterative learning control algorithm, an active iterative learning control algorithm is proposed. Different from the updating method in traditional iterative learning control algorithm, in this paper, the control input of the current iteration is given by using K-means algorithm and support vector machine (SVM) algorithm to pick out the closest state control input from the historical database. The control algorithm is verified by ethanol fermentation process. The simulation result shows that the active iterative learning control scheme based on data similarity can greatly improve the convergence speed of the system compared with the traditional one. It is worth noting that the historical data is used in the update process, so it will not affect the convergence and stability of the system, and it has good popularization value.