Data-Driven Model Predictive Control With Stability and Robustness Guarantees

We propose a robust data-driven model predictive control (MPC) scheme to linear time-invariant (LTI) systems. The uses an implicit description based on behavioral systems theory and past measured trajectories. In particular, it does not require any prior identification step, but only initially input-output trajectory as well upper bound the order of unknown system. First, we prove exponential stability nominal MPC with terminal equality constraints in case no measurement noise. For bounded additive output noise, modification scheme, including slack variable regularization cost. that application this multi-step fashion leads practical closed loop w.r.t. noise level. presented results provide first (theoretical) analysis closed-loop properties, resulting from simple, purely scheme.