Identifying Bluetooth Low Energy Devices

Physical-layer identification using hardware imperfections, known as radiometric fingerprinting, has existed for some time, but little focus has been put on Bluetooth Low Energy (BLE). This work systematically explores features for physical-layer identification of BLE. We evaluate the fingerprinting performance on different feature sets, and we discuss the potential issues with the robustness that may arise in a practical environment. Accuracy results are achieved in excess of 99%, showing potential for the system to work in security settings to safeguard a Bluetooth network.