Support for Verifying Pervasive Behavior by Mapping Task Models to Petri Nets

Abstract

Task Models (TMs) that can adapt to context are a state-of-the-art executable modeling language that has proven to be successful in the automation and support of daily user tasks in pervasive systems. TMs are intuitive and easy to use to design such systems, however, they do not provide any verification technique for ensuring their correctness. To enable verification checking, we investigate and define mappings that can translate TMs to Context-Adaptive Petri Nets (CAPNs). CAPN is a recently proposed formalism to support Petri Nets (PNs) that take into account their execution context. Using PN as a base, CAPN provides powerful techniques for behavior simulation and verification. By applying the defined mappings, pervasive system's behavior can be intuitively represented using task models and then translated to their equivalent CAPNs. These CAPNs can be then used to perform an exhaustive checking of the represented behavior at design time in order to ensure a proper and safe system execution at runtime.