Occupancy detection is becoming more popular with the growing demand for energy saving in homes and work places. Ultra wideband (UWB) radar is an attractive solution for such applications because of its low power non-ionizing emissions, low cost, high spatial resolution and ability to penetrate solid objects. In this paper, a principal component analysis-based (PCA) solution to identify the occupancy of a room using UWB radar is proposed. Preliminary tests were performed with up to 2 subjects within the detection range of the radar. The algorithm was able to determine occupancy with 100% accuracy. Furthermore, the PCA-based solution was able to determine the number of persons in the room. When one and two persons were occupying the room, PCA-based solution detected accurately the number of occupants in the room with an accuracy of 81% and 83%. The algorithm is able to distinguish two people separated by only 0.8m with 86% accuracy.

Additional Metadata
Persistent URL dx.doi.org/10.1109/MWSCAS.2017.8053237
Conference 60th IEEE International Midwest Symposium on Circuits and Systems, MWSCAS 2017
Citation
Baird, Z. (Zach), Gunasekara, I. (Isuru), Bolic, M. (Miodrag), & Rajan, S. (2017). Principal component analysis-based occupancy detection with ultra wideband radar. In Midwest Symposium on Circuits and Systems (pp. 1573–1576). doi:10.1109/MWSCAS.2017.8053237