An augmented multiple model strategy for disturbance estimation and control

Classical model-based control strategies assume a single disturbance model. In practice, the type of disturbance is often unknown, can change with time, or multiple different disturbance types can occur simultaneously. In this paper a multiple model predictive control strategy is developed to handle different disturbances, including multiple disturbances occurring simultaneously. A detailed discussion of disturbance model bank generation, state estimation and disturbance model weighting is provided, and an unconstrained multiple model predictive control solution is formulated. Simulation results demonstrate successful estimation and control of single and multiple simultaneous disturbances.