Reaching Activities by Places in the Context-Aware Environments Using Software Sensors

Context-aware systems sense the user’s situation and provide an interaction that adapts and behave accordingly. To reach a level of user’s satisfaction the systems need to be aware of activity context (i.e. what user is doing?). To sense the activity context the popular methods are hardware sensors for example tracking devices etc. or recordings of the behaviour which occurred in the past. However, as we use web services in our day to day life to get our tasks done, for example navigation, calendar, email, profile etc., stimulates the research to sense the user’s information using software sensors which can also be most relevant and circumstantial. The challenge is not only to sense the most relevant data from exchanges of messages between a user and service providers using software sensors but also to provide this data a suitable model that helps to process and store that data efficiently in both generic and more specific form. To achieve this an architecture is also needed to gather context information from software sensors. We have proposed a model that generalizes the concept according to the literature and which covers the generality in a sense which can further be lowered down to the specific domain. This paper focuses on our work regarding a novel generic semantic model for a service-based context-aware computing with emphasis on places’ information that helps systems to detect the activity of the user. It provides general taxonomy and further extends to properties and relations specific to our context-aware scenario. We have also provided an architecture that utilizes software sensors to capture user’s context, in which this model is embedded. In the final an application to a case study rounds the paper off.