A state-dependent updating period for certified real-time Model predictive control
Résumé
In this paper, a state-dependent control updating period strategy is proposed for use in interrupted implementation of real-time Model Predictive Control (MPC). The strategy can be used as soon as a certification bound is available for the underlying optimization algorithm. Moreover, a new fast-Gradient based certifiable algorithm is proposed with the associated certification bounds for convex generally constrained optimization problems.