Occupancy estimation using environmental parameters

Abstract

Occupancy information is important for energy efficient operations of Air Conditioning and Mechanical Ventilation (ACMV) systems. To predict the occupancy in a room, environmental sensors are increasingly used as cost effective and non-intrusive ways to obtain the occupancy information. This requires extraction of environmental parameters such as carbon dioxide, temperature and humidity which are used to provide a non-intrusive representation of occupancy in the room. These parameters are extracted by a data collection system consisting of a temperature sensor, humidity sensor and a gas sensor. This paper shows prediction using Logistic Regression to predict the occupancy of the room whether it is occupied or not occupied. Experiments are done in a small room in a building with few occupants.