OccuCon: A Context-aware Environment Sensing Approach Towards Indoor Occupancy Estimation

High occupancy in a room with improper ventilation and outdoor air pollution can significantly impact the occupants’ comfort and productivity which may also lead to various health-related issues. In this work, we have studied the relationship of occupancy in a classroom with the various environmental parameters prevailing in that room, such as concentration of CO2, PM1, PM2.5, PM10, and levels of Temperature, Relative Humidity, and Acoustics. All the above parameters affect indoor occupancy significantly. The real-time data acquisition has been done using a custom device. Collected contextual information helped us to find the relationships among primary features needed to develop estimation models for accurate classroom occupancy prediction. For random sampling, occupancy estimation accuracy ranges from 91% to 99%. Finally, results show that multiple environmental sensor data performed well in predicting occupancy levels.