Robust model predictive control with time-varying tubes

Abstract

© 2017, Institute of Control, Robotics and Systems and The Korean Institute of Electrical Engineers and Springer-Verlag Berlin Heidelberg. This paper focuses on the problem of robustly stabilizing uncertain discrete-time systems subject to bounded disturbances. The proposed tube-based model predictive controller ensures that all possible realizations of the state trajectory lie in the time-varying tubes so robust stability and satisfaction of the state and input constraints are guaranteed. The time-varying tubes are computed off-line so the on-line computational time is tractable. At each sampling time, the precomputed time-varying tubes are included in the optimal control problem as the constraints in the prediction horizon and only a quadratic programming problem is solved. In comparison to the algorithm that calculates the time-varying tubes on-line, the proposed algorithm can achieve the same level of control performance while the on-line computational time is greatly reduced.