Research Projects

Probabilistic Reachability and Control of Partially Observed Hybrid Systems

Jerry Ding and Claire Tomlin

Air Force Office of Scientific Research

In much of the work related to reachability analysis and controller synthesis for hybrid systems, the standing assumption tend to be that we have full state information at each time step. As is often the case in actual applications, the state measurements are subject to sensor noise and environmental disturbances, and some of the states may not even be measured. When we consider a stochastic optimal control problem in this context, this raises issues of coupling between control and state estimation. In this work, we consider these issues in the context of the probabilistic reachability problem for discrete time stochastic hybrid systems (DTSHS)[1]. Some preliminary work on switched linear Gaussian systems is currently underway.

[1]
A. Abate, M. Prandini, J. Lygeros, and S. Sastry, “Probabilistic reachability and safety for controlled discrete time stochastic hybrid systems,” Automatica, vol. 44, no. 11, pp. 2724 – 2734, 2008.