Semantic management of multiple contexts in a pervasive computing framework

► Modeled relationships among contexts using a semantic network for efficient inference. ► Integrated two different modeling approaches: knowledge-driven and data-driven. ► The semantic network (context manager) was designed based on the domain knowledge. ► Dynamic Bayesian networks (recognizer) were trained with a data-driven approach. ► Implemented a pervasive system with multi-sensors and used it to verify our method.