ILC-based adaptive minimum entropy control for general stochastic systems using neural networks

In this paper a new method for adaptive control of the general stochastic systems has been proposed. The method applies the minimum entropy control scheme to decrease the closed loop randomness of the output in an iterative learning control (ILC) basis. Both modeling and control of the plant are performed using dynamic neural networks. For this purpose, the whole control horizon is divided into a certain number of time domain sub-intervals called Batches and a P-type ILC law is employed to train the model and controller parameters so that the closed-loop tracking error is decreased batch-by- batch. The method has the advantage of decreasing the output uncertainty versus the advances of batches along the time horizon. The analysis on the proposed ILC convergence is made and demonstrable simulation results are also provided to show the effectiveness of the obtained control algorithm.