Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Bluetooth Fingerprint Identification Under Domain Shift Through Transient Phase Derivative

Haytham Albousayri, Bechir Hamdaoui, Weng-Keen Wong, Nora Basha

TL;DR

This work tackles BLE RF fingerprinting under domain shift caused by time, environment, channel, and receiver variability, with BLE’s intrinsic frequency hopping complicating identification. It introduces Transient Phase Derivative (TPD), a lightweight feature derived from the derivative of the instantaneous phase in the transient/preamble portion of frames, which suppresses channel effects and boosts device-specific signatures. The approach is validated on a 31-device BLE dataset across varied environments, channels, and receivers, showing up to 58% improvement across environments and up to 80% across receivers over baselines, and demonstrating robustness to frequency hopping. Overall, TPD enables scalable, robust RFFP in practical BLE security deployments, with public data and code to support reproducibility.

Abstract

Deep learning-based radio frequency fingerprinting (RFFP) has become an enabling physical-layer security technology, allowing device identification and authentication through received RF signals. This technology, however, faces significant challenges when it comes to adapting to domain variations, such as time, location, environment, receiver and channel. For Bluetooth Low Energy (BLE) devices, addressing these challenges is particularly crucial due to the BLE protocol's frequency-hopping nature. In this work, and for the first time, we investigated the frequency hopping effect on RFFP of BLE devices, and proposed a novel, low-cost, domain-adaptive feature extraction method. Our approach improves the classification accuracy by up to 58\% across environments and up to 80\% across receivers compared to existing benchmarks.

Bluetooth Fingerprint Identification Under Domain Shift Through Transient Phase Derivative

TL;DR

This work tackles BLE RF fingerprinting under domain shift caused by time, environment, channel, and receiver variability, with BLE’s intrinsic frequency hopping complicating identification. It introduces Transient Phase Derivative (TPD), a lightweight feature derived from the derivative of the instantaneous phase in the transient/preamble portion of frames, which suppresses channel effects and boosts device-specific signatures. The approach is validated on a 31-device BLE dataset across varied environments, channels, and receivers, showing up to 58% improvement across environments and up to 80% across receivers over baselines, and demonstrating robustness to frequency hopping. Overall, TPD enables scalable, robust RFFP in practical BLE security deployments, with public data and code to support reproducibility.

Abstract

Deep learning-based radio frequency fingerprinting (RFFP) has become an enabling physical-layer security technology, allowing device identification and authentication through received RF signals. This technology, however, faces significant challenges when it comes to adapting to domain variations, such as time, location, environment, receiver and channel. For Bluetooth Low Energy (BLE) devices, addressing these challenges is particularly crucial due to the BLE protocol's frequency-hopping nature. In this work, and for the first time, we investigated the frequency hopping effect on RFFP of BLE devices, and proposed a novel, low-cost, domain-adaptive feature extraction method. Our approach improves the classification accuracy by up to 58\% across environments and up to 80\% across receivers compared to existing benchmarks.

Paper Structure

This paper contains 30 sections, 3 equations, 11 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Key Contributions
  4. BLE Hardware Impairments
  5. Ideal GFSK Modulation
  6. Impaired GFSK Modulation
  7. TPD: the Feature Extraction Approach
  8. Discrete Time Domain TPD
  9. Motivation Behind Using TPD Instead of Raw IQ
  10. TPD's Effectiveness in Capturing BLE Hardware Impairments
  11. Carrier Frequency Offset $(f_{CFO})$
  12. IQ Imbalance ($IQ_{Amp}$ and $IQ_{Phase}$)
  13. IQ Offset ($I_{DC}$ and $Q_{DC}$)
  14. Maximum (Peak) Frequency Deviation ($f_m$)
  15. Bandwidth Duration (BT) Product
  16. ...and 15 more sections

Figures (11)

  • Figure 1: Testing classification accuracy for 31 BLE devices across different environments and different receivers. More on this experiment/testbed can be found in Sec. \ref{['sec:Testbed']}
  • Figure 2: TPD representation extracted from signals sent by different devices across different channels in the same wired setup
  • Figure 3: The block diagram of the proposed technique.
  • Figure 4: TPD representation under the effect of hardware impairments
  • Figure 5: The impact of $\theta_{PO}$ on TPD and Raw IQ
  • ...and 6 more figures