Robust Multi-Model Predictive Control via Integral Sliding Modes

This paper presents a novel optimal control approach for systems represented by multi-model, i.e., finite set of models, each one corresponding to different operating point. The proposed scheme is based on the combined use model predictive (MPC) and first order integral sliding mode control. component plays important role rejecting matched uncertainty terms possibly affecting plant, thus making controlled equivalent system behave as nominal multi-model. A min-max multi-model MPC problem solved using without further robustness oriented add-ons. In addition, design performed so keep computational complexity limited, facilitating practical applicability proposal. Simulation results show effectiveness approach.