An Integral-Model Predictive Controller with Finite Memory for Trajectory Tracking

Abstract

In this paper, an integral-model predictive control (i-MPC) scheme with finite-memory was proposed to track a time-varying signal. It has been shown that with the use of the so-called i-MPC, the persistent steady-state error can be made smaller. In order to investigate its performance, the so-called i-MPC was used to steer a robot along a reference path. It has been shown that time-varying signal tracking performance and convergence characteristics of the so-called i-MPC scheme is better than that of a regular model predictive control and a regular model predictive control with an integral action.