Formal verification of context and situation models in pervasive computing

Pervasive computing is a paradigm that focuses on the availability of computer resources anytime anywhere for any application and supports non-intrusive integration of computing services into everyday life. Context awareness is the core feature of pervasive computing. High-level context awareness can be enhanced by situation awareness that represents the ability to detect and reason about the real-life situations. In this article we propose, analyze and validate the formal verification method for situation definitions and demonstrate its feasibility and efficiency. Situations are often defined manually by domain experts and are, therefore, susceptible to definition inconsistencies and possible errors, which in turn can cause situation reasoning problems. The proposed method takes as an input properties of situations and dependencies among them as well as situation definitions in terms of low-level context features, and then either formally proves that the definitions do comply with the expected properties, or provides a complete set of counterexamples — context parameters that prove situation inconsistency. Evaluation and complexity analysis of the proposed approach are also presented and discussed. Examples and evaluation results demonstrate that the proposed approach can be used to verify real-life situation definitions, and detect non-obvious errors in situation specifications.