Distributionally Robust Model Predictive Control With Total Variation Distance

This paper studies the problem of distributionally robust model predictive control (MPC) using total variation distance ambiguity sets. For a discrete-time linear system with additive disturbances, we provide conditional value-at-risk reformulation MPC optimization that is in expected cost and chance constraints. The constraint over-approximated as simpler, tightened reduces computational burden. Numerical experiments support our results on probabilistic guarantees efficiency.