Occupancy matters: toward an occupancy-driven ventilation system using WiFi infrastructure and neural network

Buildings accounted for 20% of total energy consumption and 54% of electricity usage in 2013 in Canada. Heating Ventilation and Air conditioning system is the main consumer of energy in the buildings. The common approach for designing a ventilation system is a predefined schedule based on the maximum capacity disregard the actual number of occupants. We believe that passive use of already existing WiFi infrastructures can replace the monitoring sensors and cut the cost of energy and extra sensors installation. A field study was conducted in graduate offices of Ryerson University to examine the opportunity of energy saving by changing the fixed ventilation schedule to the occupancy driven one. The number of occupants had been determined using pre-existing WiFi infrastructure and by using the real time occupancy data, the new system achieved 76% reduction in ventilation energy consumption. To further investigate the potentials of WiFi infrastructure, Finger Printing and Neural Network method had been used to map the occupant’s location by analyzing the Received Signal Strength Indicator (RSSI) of the wifi equipped device. The results showed 95% accuracy in the first round of testing and 92% accuracy after 1 week of retesting the model by using pattern recognition technique. Employing this approach could lead to even more energy saving by assigning the required airflow to each subzone proportionally to the number of its occupants.