Model Predictive Control with Both States and Input Delays

Abstract

This study investigates the problem of model predictive control (MPC) for active suspension systems with both states and input delays. The model uncertainty is assumed to be polytopic, and sufficient conditions are proposed in terms of linear-matrix inequalities (LMIs), which can be easily solved by an efficient convex optimization algorithm. The problem of minimizing an upper bound on the ‘worst-case’ performance objective function is reduced to a convex optimization involving linear matrix inequalities (LMIs). At each time set, a feasible state feedback is obtained by minimizing an upper bound of the ‘worst-case’ quadratic objective function over on infinite horizon subject on constraints on inputs. Finally, a quarter-vehicle model is exploited to demonstrate the effectiveness of the proposed method.