Situation-based testing for pervasive computing environments

A challenge for designing pervasive computing systems, particularly for indoor sensor-rich environments, is the complexity of causal relationships between contextual inputs, exhibited system behaviour, and overall appropriateness of resulting outcomes. Two key challenges faced by designers when testing these systems lie in the difficulty of monitoring the deployment environment for inappropriate outcomes and subsequently tracing the physical and digital causal factors leading to specific situations. This paper presents InSitu, a situation-based testing approach that applies generalised situation specifications to a global representation of simulated environment state to detect instances of specified situations. An access control case study demonstrates the modelling capabilities of InSitu and notes the importance of the overlap, exclusivity, and subsumption relationships between situation specifications in order to achieve informative results.