Adaptive Prescribed Performance Tuning for Model Predictive Control

Abstract

A novel adaptive prescribed performance model predictive control (APPMPC) scheme for inductively coupled plasma (ICP) generation system is developed in this paper. First, the ICP generation control system is introduced, which is followed by a general MPC controller. Then, an APPMPC scheme is proposed, including the design of cost function and tuning for the prescribed performance function. The cost function of APPMPC could be obtained by transforming the cost function of classical constrained MPC based on the prescribed performance function and error transmission theory. The prescribed performance function parameters of the APPMPC could be tuned automatically with predefined overshoot and steady-state error, which enables APPMPC to be capable of guaranteeing the desired transient and steady-state prescribed performance of the ICP generation control system. Moreover, the proposed APPMPC converts the constrained quadratic programming (QP) problems of classical constrained MPC to unconstrained QP problems, thus significantly reducing computational costs. The simulation results show that the APPMPC method can guarantee the prescribed performance of tracking error, overshoot, steady error, less computational burden, and uniformly bounded output for the ICP generation system.