Context Awareness Computing in Smart Spaces Using Stochastic Analysis of Sensor Data

In building a smart space, it becomes more critical to develop a recognition system which enables to be aware of contexts, since the appropriate services can be provided under the accurate recognition. As services satisfying for desires of individual human residents are more demanding, the necessity for more sophisticated recognition algorithms is increasing. This paper proposes an approach to discover the current context by stochastically analyzing data obtained from sensors deployed in the smart space. The approach proceeds in two phases, which is to build context models and to find one context model matching the current state space, however we mainly focus on the phase building context models. Experimental validation supports the approach and approved validity.