Robust dual adaptive MPC for output tracking of linear systems

In this work, we present a novel robust dual adaptive model predictive control scheme for linear discrete-time systems with parametric uncertainty in the output and affected by a state-disturbance and a measurement noise. The proposed MPC framework promotes learning of the unknown parameters and in the meantime tracks a desired target output despite the presence of a state disturbance. This is possible through a suitable restriction of the state, input and output constraints that covers the whole possible range of the state disturbance. We prove that the proposed controller is practically stable and we ensure robust constraint satisfaction for state, input and output, also in the presence of parametric uncertainty and bounded noises. We corroborate the theoretical result through a numerical example simulated on Matlab.