Real-time optimal feedback control of switched autonomous systems

This paper presents a novel optimal adaptive feedback control concept for nonlinear switched autonomous systems. Our proposed controller implements a closed-loop switching law, which is represented by a parametrized switching plane in the state space. During the plant operation, the plane parameters are incrementally adapted to state measurements in real time by iteratively solving an associated nonlinear optimization problem over a finite time horizon. The combination of this control approach with ideas from model predictive control yields the so-called crawling window optimal control scheme that achieves the optimal stabilization of periodically operated switched systems.