Machine Learning applied to Bluetooth Low Energy Beacons Indoor Location
Bluetooth Low Energy (BLE) beacons can be used to enable the physical web, proximity, or fine tuned indoor location information. We focus on the indoor micro-location or locating people or objects with sub metre accuracy indoors.
Who are we?
- Mae Kennedy - Masters Student, BASc, University of Guelph, School of Engineering
- Graham W. Taylor - Associate Professor, University of Guelph, School of Engineering
- Petros Spachos - Assistant Professor, University of Guelph, School of Engineering
We are interested in the following aspects of BLE beacons:
- Security
- Performance
- Usability
Our works under this project include:
- Beaconpi a system for aggregating beacon data collected by Raspberry Pi units, being actively developed for
- BLE Beacon Based Patient Tracking in Smart Care Facilities a paper describing our system for indoor location with BLE Beacons
- Privacy and Security Concerns with BLE Beacons for Healthcare: a submitted paper the addresses the privacy and security implications of using BLE beacons for indoor location in the healthcare vertical
Future work:
- Utilize bayesian filtering to accomplish sub metre accuracy in indoor location
- Utilize convolutional neural networks to determine non line of sight conditions and predict the true distance based on this information