Research output: Contribution in Book/Report/Proceedings - With ISBN/ISSN › Conference contribution/Paper › peer-review
Research output: Contribution in Book/Report/Proceedings - With ISBN/ISSN › Conference contribution/Paper › peer-review
}
TY - GEN
T1 - Creating a room connectivity graph of a building from per-room sensor units
AU - Ellis, Carl
AU - Scott, James
AU - Constandache , Ionut
AU - Hazas, Michael
PY - 2012/11
Y1 - 2012/11
N2 - Sensor and actuator networks are often installed in buildings for energy-related applications such as lighting and climate control. Such systems require metadata about the deployed hardware (e.g. which room each is in, what the function of each room is) in order to operate effectively. In this paper we present methods to automatically determine such metadata, in particular the room connectivity graph (i.e., which rooms share a doorway/interior window). Crucially, our method works with just one sensor unit per room, does not require special placement of any of the sensors, and can therefore work on data from existing widely-deployed applications (such as burglar alarms). We apply this method to a 30-day data set from single per-room sensor units deployed in two residential homes in the United Kingdom. Room connectivity is determined based on: spillover of artificial light between rooms; occupancy detections due to movement between rooms; and a fusion of the two. The fusion of both techniques is shown to work better than either technique alone, with a 93% true positive rate and 0.5% false positive rate (aggregate across both houses), and a convergence time of under a week.
AB - Sensor and actuator networks are often installed in buildings for energy-related applications such as lighting and climate control. Such systems require metadata about the deployed hardware (e.g. which room each is in, what the function of each room is) in order to operate effectively. In this paper we present methods to automatically determine such metadata, in particular the room connectivity graph (i.e., which rooms share a doorway/interior window). Crucially, our method works with just one sensor unit per room, does not require special placement of any of the sensors, and can therefore work on data from existing widely-deployed applications (such as burglar alarms). We apply this method to a 30-day data set from single per-room sensor units deployed in two residential homes in the United Kingdom. Room connectivity is determined based on: spillover of artificial light between rooms; occupancy detections due to movement between rooms; and a fusion of the two. The fusion of both techniques is shown to work better than either technique alone, with a 93% true positive rate and 0.5% false positive rate (aggregate across both houses), and a convergence time of under a week.
U2 - 10.1145/2422531.2422563
DO - 10.1145/2422531.2422563
M3 - Conference contribution/Paper
SN - 978-1-4503-1170-0
SP - 177
EP - 183
BT - BuildSys '12 Proceedings of the Fourth ACM Workshop on Embedded Sensing Systems for Energy-Efficiency in Buildings
PB - ACM
CY - New York
ER -