Model-Predictive Control for Singular Semi-Markov Jump Systems With Application to Hybrid Terrestrial/Aerial Bicopters

This paper is concerned with the issue of model predictive control (MPC) for a class of stochastic hybrid singular systems, where a semi-Markov kernel orchestrates the switching among the system modes with bounded sojourn time. An MPC framework is proposed accounting for singular subsystems, a more common case in actual application for the first time, which departs from conventional assumptions of mode-independent subsystem dimensions. By virtue of the receding horizon optimization, infinite horizon model predictive controllers with the recursive feasibility and stochastic stability guarantee are obtained. The necessity and validity of the proposed model predictive controllers are demon strated using illustrative examples involving hybrid terrestrial/aerial bicopters.