Situvis: A Visual Tool for Modeling a User’s Behaviour Patterns in a Pervasive Environment

Abstract

Abstract. One of the key challenges faced when developing contextaware pervasive systems is to capture the set of inputs that we want a system to adapt to. Arbitrarily specifying ranges of sensor values to respond to will lead to incompleteness of the specification, and may also result in conflicts, when multiple incompatible adaptations may be triggered by a single user action. We posit that the ideal approach combines the use of past traces of real, annotated context data with the ability for a system designer or user to go in and interactively modify the specification of the set of inputs a particular adaptation should be responsive to. We introduce Situvis, an interactive visualisation tool we have developed which assists users and developers of context-aware pervasive systems by visually representing the conditions that need to be present for a situation to be triggered in terms of the real-world context that is being recorded, and allows the user to visually inspect these properties, evaluate their correctness, and change them as required. This tool provides the means to understand the scope of any adaptation defined in the system, and intuitively resolve conflicts inherent in the specification.

Internet of things: Vision, applications and research challenges

“…Services in IoT are expected to be able to seamlessly adapt to different situations and contexts. A number of research efforts for building self-adaptive situated services have been undertaken in the last few years [37][38][39][40][41][42]. However, we are still far from reaching a global understanding of how to develop self-adaptive services presenting the flexibility level required by IoT scenarios.…”

Ambient Middleware for Context-Awareness (AMiCA)

“…This information is often unreliable due to the inaccuracy and lack of precision of sensors interpreting the environment (Dobson et al, 2007). Current research also looks at hybrid approaches which incorporate the use of semantic web technologies such as ontology's for modelling and acquiring context (Clear et al, 2007).…”