Smart Building Energy Management Aspects of Human Mobility Detection and Tracking

  • Haleimah A. Zeyoudi

Student thesis: Master's Thesis


In a typical open office building, approximately 30 to 40 percent of energy is consumed in systems such as air conditioning, ventilation, lighting and others. Nowadays most modern air conditioning systems in buildings do not take into account actual human occupancy and activities which is inefficient and creates energy waste. Therefore, in order to achieve an efficient air conditioning usage, we need to take into account the mobility behaviour of occupancy in each zone of a building, which allows us to intelligently tune the air conditioning system according to the state of occupancy without compromising the comfortability. We deployed a video capturing and processing system to acquire temporal occupancy information. To this end, we developed occupancy counting software based on-Microsoft Kinect sensor platform. Our software system is capable of detecting and tracking human activity inside a building. One of the drawbacks of the current air conditioning system is the inability to adapt to the changing occupancy and activities instantaneously. Therefore, the management and control of an intelligent HVAC system should be based on the dynamic occupancy data. In our research, we proposed an approach for energy savings that integrates a real-time occupancy data with building management systems. This approach leads us to the creation of an occupancy monitoring and air conditioning control system based on dynamic occupancy counting. This system is able not only to track mobility of occupants but also to show the status of occupancy in different zones of a building. Based on the prediction of future occupancy level of particular zones in the building, an intelligent system can adjust air conditioning parameters optimally. In this thesis, we introduced a new concept on human mobility detection and tracking in real-time, based on occupancy counter software. We provided the first study that uses multiple Kinect sensor as a sensor in building monitoring and management. It shows how Kinect could be applied successfully for the fields of detection and tracking. We proposed an occupancy model that display's a visualized data and predicate a future occupancy through a Markov Model transition matrix. Through extensive real time testing of our system, the experimental results have shown a potential energy saving around 22.1% by applying a customized occupancy schedule and customized set point depending on occupancy density.
Date of AwardMay 2014
Original languageAmerican English
SupervisorChi Kin Chau (Supervisor)


  • Mobility
  • Management Energy
  • Consumption
  • Energy Waste
  • Air Conditioning System
  • Microsoft Kinect Sensor
  • HVAC System
  • Markov Model.

Cite this