Control input synthesis for hybrid systems using informed search

We present a method for synthesizing a sequence of robust control inputs for a class of hybrid systems. Our goal is the generation of a control sequence that drives the system from a given initial state set to a pre-specified goal set without violating constraints on the system state, under the assumption that the hybrid system is exposed to bounded disturbances. We use a technique that combines dynamic programming and informed search. The control sequence generated by our synthesis procedure is guaranteed to meet safety requirements. An extension to nonlinear systems is presented and computational time is compared to a mixed-integer programming approach for computing an optimal but non-robust solution to the problem