Dealing with uncertainty in context-aware mobile applications

The exploitation of context-awareness, especially in mobile devices bears a huge potential. For example, mobile workers benefit from systems that adapt security settings to the current situation. However, context-aware computing strongly relies on raw data from various sources that might be neither trustworthy nor authoritative. In this work, we present a context model that explicitly reflects security and relevance of context information sources in order to improve context detection. We introduce a security rating denoting the trustworthiness of the context information, i.e., its vulnerability to forgery, and a relevance rating denoting the source’s decisive impact on context detection