RAY: Batteryless Occupancy Monitoring Using Reflected Ambient Light
Keywords: Occupancy, Batteryless, Intermittent, Energy harvesting, Energy as Data
TL;DR: Ray is a batteryless, doorframe/passageway-mounted room-level occupancy monitoring sensor that uses changes in indoor ambient light reflections to detect people entering and exiting a room or hallway and estimate their direction of travel.
Abstract: Room-level occupancy-tracking systems enable intelligent control of building functions like air conditioning and power delivery to adapt to the needs of their occupants. Unfortunately, existing occupancy-tracking systems are bulky, have short battery lifetimes, are not privacy-preserving, or only provide coarse-grain occupancy information. Furthermore, retrofitting existing infrastructures with wired sensors is prohibitively expensive.
In this paper, we present Ray, a batteryless, doorframe/passageway-mounted room-level occupancy monitoring sensor that uses changes in indoor ambient light reflections to detect people entering and exiting a room or hallway and estimate direction of travel. We evaluated Ray in mixed lighting conditions on both sides of the doorway in an office-style setting, using subjects with a wide variety of physical characteristics. We conducted 881 controlled experiments in 7 doorways with 9 individuals and achieved a total detection accuracy of 100% and movement direction accuracy averaging 96.4%. Furthermore, we deployed Ray sensors for 64 days in 5 locations, comparing them with a commercial batteryless occupancy sensor. Ray outperformed the commercial sensor, particularly where traffic is moderate to heavy. Ray demonstrates that ambient light reflections provide both a promising low-cost, long-term sustainable option for monitoring how people use buildings and an exciting new research direction for batteryless computing.
Area: Wireless Embedded Systems
Type: Solution
Conflicts: Josiah Hester, Pat Pannuto
Potential Reviewers: Shijia Pan, Bernd-Christian Renner
Revision: No
Contact Email: TobiasRN@Wofford.edu
Submission Number: 3
Loading