doiSerbia

System design for passive human detection using principal components of the signal strength space

Mrazovac Bojan (Faculty of Technical Sciences, Novi Sad)
Bjelica Milan Z. (Faculty of Technical Sciences, Novi Sad)
Kukolj Dragan (Faculty of Technical Sciences, Novi Sad)
Todorović Branislav M. (RT-RK, Institute for Computer Based Systems, Novi Sad)
Vukosavljev Saša (RT-RK, Institute for Computer Based Systems, Novi Sad)

In this article, device-free human presence detection method based on principal components analysis of the radio signal strength variations is proposed. The method increases user awareness for automated power management interaction in residential power saving systems. Motivation comes from a need for decreasing the installation complexity and development costs induced by the integration of specific human presence detection sensors. The method exploits the fact that a human body interferes with radio signals by introducing irregularities in the radio signature which indicate possible human presence. By analyzing radio signals between radio transceivers embedded in 2.4 GHz wireless power outlets, the original environment is not visually modified and a certain level of sensorial intelligence is introduced without additional sensors. The analysis of the signal strength variations in principal components space enhances the detection accuracy level of human presence detection method, retaining low installation costs and improving overall energy conservation.

Keywords: energy awareness, human presence detection, principal components analysis, radio irregularity, RSSI, smart outlets, Zigbee

Projekat Ministarstva nauke Republike Srbije, br. TR-32029 and by the Secretary of Science and Technology Development of the Province of Vojvodina under Grant 114-451-2434/2011-03