SIGNATURE WORK
CONFERENCE & EXHIBITION 2022

Team rooms and meeting rooms are popular places for DKU students and faculty to stay in. Their occupancy is usually demand-based and not regular. Oftentimes, to find an empty room with no one occupied, people need to go to different buildings and different floors to check by themselves, which would usually spend much time and probably found nothing. Therefore, the detection for occupancy has many practical meanings. The aim of this project is to propose a noninvasive method for room occupancy detection. In this project, a systematic review of the current indoor occupancy detection is presented and multiple methods based on ventilation, temperature, WI-FI, etc. are compared. This project also proposes a CO2 concentration-based method to identify the occupancy. An ordinary differential equation is adopted to model the variation of CO2 concentration in a DKU team room with some unknown parameters, like human expiration rate and room size. These parameters are estimated by fitting with the collected data. After the environment parameters are estimated, our model is used to estimate the occupancy with new test data. Two denoising methods, Fast Fourier Transform and Savitzky-Golay filter are also compared in data preprocessing.