Generalized regression neural networks-based data-driven iterative learning control for nonlinear non-affine discrete-time systems

In this paper, a novel neural-network iterative learning control (IL, ILC) protocol for generalized regression neural networks (GRNN) based on global space–time is proposed for a class of unknown nonlinear discrete-time systems. The proposed controller is an ideal controller with multidimensional gain based on the global space–time in the iteration domain. The control gain is automatically learned by updating the matrix of the GRNN with correction factors based on the accumulation of global system input and output (I/O) data along the iteration axis. By a rigorous theoretical analysis, properties that can drive the tracking error to be uniformly bounded in the iteration axis are established for the proposed control protocol with the goal of achieving optimal control. The effectiveness and adaptability of the proposed control protocol have been validated through numerical simulations and two comparative examples on general high-frequency signal tracking.