Multi-rate model predictive control to reject periodic disturbance with reduced ripple

Repetitive model predictive control is an effective method to track a periodic signal as well as reject a periodic disturbance. However, since repetitive controller uses discrete-time model, there always exists ripple when controlling a continuous system with periodic disturbance. This paper proposes a new way for repetitive model predictive control by using multi-rate model. Due to the multi-rate model, the model predictive controller can optimize the system outputs during the period between two sampling times, and then provide a more delicate control action, which can result in the reduced ripple. The simulation demonstrates its effectiveness.