Device-Free Human Motion Detection Using Single Link WiFi Channel Measurements for Building Energy Management

Human motion is a vital parameter signaling indoor occupancy in occupant-centric building energy management systems. In this letter, a WiFi-based device-free sensing model employing ESP32—a low-cost embedded Internet of Things device, is proposed as a suitable retrofit for human motion sensing in new and existing buildings. A low complexity feature set derived from WiFi received signal strength indicator and channel state information samples collected by ESP32 is utilized for training ensemble machine learning models. The proposed model shows an accuracy of up to 99.4% in classifying human motion in three different rooms of a residential building, with a single WiFi access point. The proposed model can be employed for motion-driven smart load control and for a typical residential occupancy scenario, this approach exhibits a potential monthly electricity savings of up to 66.26% and a three times reduction in CO2 emissions.