Robust Radio Frequency Fingerprint Identification for Bluetooth Low Energy under Low SNR and Channel Variations

Radio frequency fingerprint identification (RFFI) is a promising technique for authenticating Internet of Things (IoT) devices by leveraging unique RF hardware impairments. However, RFFI is vulnerable to channel variations and low signal-to-noise ratio (SNR) conditions. In this paper, we proposed a robust RFFI system specifically designed to tackle these issues for Bluetooth Low Energy (BLE), which is a popular IoT technology. Our system integrated a denoising autoencoder (DAE) to enhance feature robustness under low SNR conditions and employed data augmentation to mitigate the impact of channel and noise effects. We created a testbed consisting of 18 commercial off-the-shelf (COTS) BLE devices and a USRP N210 software-defined radio (SDR) platform and then carried out extensive experimental evaluation under various channel conditions. The experiments involved line-of-sight (LOS) and non-line-of-sight (NLOS) propagation as well as dynamic and static channels. The results demonstrated that our approach consistently achieved over 95% accuracy in high SNR environments and maintained strong performance with over 75% accuracy at low SNR levels (10 dB).