From Signal to Signature: Enhanced BLE Device Fingerprinting with Environmental Context

Abstract:
Bluetooth Low Energy (BLE) is present in nearly every modern device that utilizes wireless communications, making it a valuable target for forensic investigations. Unfortunately, BLE addresses can change frequently, making them unsuitable for accurate device identification and attribution. Prior work has shown that BLE transmissions exhibit device-specific physicallayer artifacts, such as Carrier Frequency Offset (CFO), which can be exploited for device fingerprinting. However, these characteristics are not static and are affected by environmental factors such as temperature. In this paper, we investigate the forensic potential of BLE fingerprinting based on CFO measurements by modeling the CFO behavior as a function of environmental temperature across multiple devices. Using a software-defined radio, we collect BLE advertising packets under controlled thermal conditions to construct CFO–temperature curves that are unique to each device. Our findings demonstrate that incorporating temperature as a variable enhances the robustness of BLE device fingerprinting, enabling reliable device identification over extended periods and across varying environmental conditions. This approach strengthens the forensic capabilities by enabling persistent, passive, and non-intrusive device tracking techniques that are immune to address changes.