A modified design framework for nonlinear model predictive iterative learning control

Abstract

As advanced control strategies, both iterative learning control (ILC) and model predictive control (MPC) are widely used in industrial process. Because ILC cannot eliminate the non-repetitive disturbances, ILC and MPC are integrated as model predictive iterative learning control (MPILC) to improve the capability of rejecting disturbances. Although the typical MPILC has a good tracking performance, there is also left some aspects to be developed. Based on a fuzzy model, a modified nonlinear model predictive iterative learning control (NMPILC) is proposed to achieve a better tracking performance and speed up the learning rate. The performance of the modified NMPILC is illustrated by a PH neutralization process.