Example distributed sensor network control hierarchy

Our model uses discrete event dynamic systems (DEDS) formalisms. DEDS have discrete time andstate spaces. They are usually asynchronous and nondeterministic. Many DEDS modeling and control methodologies exist and no dominant paradigm has emerged [2]. We use Petri nets, as will be described in Section 51.2, to model the plants to be controlled. Our sensor network model has three intertwined hierarchies, which evolve independently. We derive controllers to enforce system consistency constraints across the three hierarchies. Three equivalent controllers are derived using (i) Petri net, (ii) vector addition and (iii) finite-state machine (FSM) techniques. We compare the controllers in terms of expressiveness and performance. Innovative use of Karp-Miller trees [3] allows us to derive FSM controllers for the Petri net plant model. In addition, we show how FSM controllers can be derived automatically from control specifications in the proper format.