Reachability-based diagnosability for hybrid systems

Abstract

In this work, we present the main concepts related to hybrid systems, that are dynamic systems that combine discrete and continuous behaviors. These systems can be modeled by hybrid automata or transition systems. Hybrid automata allow more wealth to the model, however, it is not always possible to predict its behavior. Depending on the type of analysis of the system, such as the verification of certain properties, a higher level of abstraction is preferred and, thus, modeled by a transition system, which presents more structured tools for checking properties. This work also presents a new definition of diagnosability combining the discrete event systems (DES) diagnosability with reachability analysis to compare continuous behaviors. Furthermore, we present a case study of fault diagnosability analysis of systems modeled by hybrid automata. This example demonstrates the advantage of performing the state reachability analysis associated with the continuous-time dynamics of the hybrid model. With this approach, it is possible to diagnose failures that would not be possible by using purely DES techniques.