Factory monitoring and control with mixed hardware/software, discrete/continuous models

Many complex systems today, such as robotic networks, automobiles and automated factories, consist of hardware components whose functionality is extended or controlled by embedded software and which exhibit continuous dynamics. We address the problem of monitoring and control in such systems with a twofold contribution. First, we extend probabilistic hierarchical constraint automata (PHCA), introduced in previous work as a means to compactly describe uncertain hardware and complex software behavior, to hybrid PHCA (HyPHCA). These allow to model continuous behavior in the form of differential equations. Continuous behavior can be conservatively approximated with discrete Markov chains, and in previous work we showed how to transform PHCA monitoring into a constraint optimization problem that can be solved using off-the-shelf reasoners. Our second contribution is to show how to combine these and additional known methods to use a HyPHCA to monitor the internal state and plan for contingencies in a rich class of mixed hardware/software, discrete/continuous systems. Preliminary results of our approach for an industrial filling station scenario demonstrate its feasibility.