Robust data-driven predictive control using reachability analysis

We present a robust data-driven control scheme for an unknown linear system model with bounded process and measurement noise. Instead of depending on in traditional predictive control, controller utilizing reachable regions is proposed. The are based matrix zonotope recursion computed only noisy input-output data trajectory the system. assume that noise contained sets. While we knowledge these bounds, no about statistical properties assumed. In noise-free case, prove presented purely results equivalent closed-loop behavior to nominal scheme. case noise, our proposed guarantees constraint satisfaction, which essential safety-critical applications. Numerical experiments show effectiveness comparison model-based schemes.