Non-Invasive Occupancy Monitoring Using Sustainable Backscatter Tags

Recently, WiFi signals have been explored for precise occupancy sensing, but they require a powered transmitter and receiver. This paper demonstrates the viability of novel backscatter occupancy sensing technology that uses ultra-high frequency radio waves transmitted by a reader to multiple passive sustainable tags. Our proposed sensing model uses sub-dollar cost passive tags that do not require active energy consumption and cover an area of interest for non-invasive and eco-friendly surveillance using backscattered tags. We use variations in the backscattered received strength indicator (RSSI) measured by the reader to train a support vector machine, resulting in a green sensing model that can yield an estimated accuracy of around 99% for counting up to three occupants. Our experimental demonstration using commodity hardware proved that our framework works well even with disturbances in the surrounding environment, and this nonintrusive technology is superior to existing WiFi sensing for occupancy monitoring in designated areas.