Computer Science Colloquia
Wednesday, April 17th, 2012
Advisor: Kamin Whitehouse
Attending Faculty: Mary Lou Soffa (Chair)
Rice Hall, Room 204, 10:00 am
Master Project Presentation
Evaluating the Effect of Sensor Placement on Automated Sensor-Map Generation in Smart Homes
According to the U.S. Department of Energy, Residential buildings accounted for 22% of total primary energy consumption in the U.S. in 2009. Smart homes are valuable to increase the energy efficiency of the Residential buildings. Within a smart home many sensors and actuators are interconnected to form a control system. However, manual configuration of these devices is difficult and time-consuming and an important barrier to adoption of the system to the general mass. Therefore in this paper, we present and compare a family of solutions that automatically generate a map of the home and the devices within it from the smart home sensors themselves, without using any additional specialized tool. The current state of the art i.e. "Smart Blueprints" system uses sensors on all the windows and doorways in the house, which is costly for average homeowners and can make the system difficult to install and use. We demonstrate the Smart Blueprints using a variety of sensor combinations and placements, including light sensors, motion sensors, and magnetometers deployed on the doors and/or windows of the home. Then, we evaluate how well each scenario can be used to map the home configuration. For evaluation, we deployed over 200 sensors in 7 different houses at different locations and compared the ability to use a variety of techniques to map out the configuration. We show that, using any of the sensor placement scenarios, this system can automatically narrow the configuration down to 2-7 candidates on average using only one week of collected data.