Extracting high-level user context from low-level mobile sensors data

Many modern smart mobile devices integrate a multitude of various sensors which can enrich the usability of today mobile applications. However, collecting and analyzing data from sensors requires increased processing, storage and energy resources on mobile devices. On the other hand, context-aware mobile applications require an estimation of high-level user context, in order to provide the user with specific services. It is not simple to model and estimate in real-time user context, even though a lot of low-level sensors data are available. Therefore, in this paper we investigate how data collected from different mobile sensors and communication interfaces can be used to extract the user context.