The Application of Sensor Networks to Lighting Control

Abstract

This paper evaluates the application of sensor networks to lighting control, and had two goals. The first goal was to evaluate the utility of using more than one sensor to detect occupancy. Sensor networks were installed in a sample of private offices and cubicle workstations. Large individual differences in sensor response to occupancy were observed in all monitored areas: the percent difference in occupied time measured by sensors in ten private offices ranged from 16.1 percent to 74.0 percent (mean 44.7 percent): the differences between detectors measuring occupancy in 23 cubicle workstations ranged from 14 percent to 91 percent (mean 53.7 percent). These data establish the considerable uncertainty in occupancy measurement that exists with current systems. Current control systems compensate for this uncertainty by applying a long time delay before switching lights off. Since a sensor network more accurately measures occupancy, shorter time delays can be applied to reduce system operating time and save energy. Applying a 5-minute time delay to sensor network data, instead of the 20 minutes that is more typical in current single sensor applications, reduced operating time by an additional 22.4 percent to 33.3 percent (relative to the reductions that would have resulted from use of a single sensor). A longer time delay of 10 minutes applied to the sensor network data stream produced reductions of 8.4 percent to 24.7 percent, relative to the reduction that would have resulted from use of a single sensor with a 20-minute time delay.

Keywords:

• Light control
• Networks
• Sensors.