Anonymous Proximity Beacons from Smartphones

Abstract

This project investigates the usage of Bluetooth-enabled smartphones for privacy-preserving human proximity detection. Smartphones can transmit identifiers which can be heard by other smartphones nearby. A naive solution where each smartphone continually transmits a fixed identifier would make it trivial for a third party to track users; this project explores a solution for Bluetooth Low Energy and Bluetooth 5 to make tracking difficult by using mutable pseudo-anonymous identifiers that can only be reversed by a de-anonymisation server. An example application is developed: the server can return location information to requesting users, emulating the functionality of static proximity beacon systems—except without the deployment costs.

Publication
Bachelor's Dissertation
Avatar
Shyam Tailor
Machine Learning PhD Student

My research interests include enabling efficient on-device machine learning algorithms through hardware-software co-design, and exploring the applications enabled by these advances.