Energy efficient communication in smart building WSN running distributed Hidden Markov chain presence detection algorithm

Occupancy detection in smart buildings suffers from high sensor unreliability. The combination of data from multiple sensors can largely improve reliability. However, battery-powered sensor nodes have communication limitations. This paper addresses the problem of presence detection in a building as Distributed Sensing of a Hidden Markov Model (DS-HMM). Optimal solutions require excessive communication over a perfectly reliable channel. We propose and investigate an efficient transmission policy, jointly with a fusion algorithm, to merge data from various HMMs running separately on all sensor nodes. The algorithm showed improved performance and potential energy savings in a real world set-up, where user presence and sensor errors may not exactly follow idealized model assumptions.