Self-tuning control design based on recursive parameter estimation and robust receding horizon control

This paper presents a new self-tuning controller for a linear system with unknown time-varying parameters, which, however, have a noise-like disturbance behavior. The regulator employs two steps. The first one is to estimate system´s parameters and their estimation error bounds, recursively, via Kalman filtering. Based on these estimated parameters and error bounds, the second one is to construct a receding horizon controller robust to such estimated error bounds. The most important contribution of this paper is that the new method successfully combines estimation and control procedures so as to ensure the asymptotic stability of the resulting closed-loop system, which has been a major issue in adaptive control