On the Domain Generalizability of RF Fingerprints Through Multifractal Dimension Representation

TL;DR

This work tackles the domain adaptation challenge in RF fingerprinting by introducing Variance Fractal Dimension Trajectory (VFDT), a multifractal representation of the RF IQ signals that captures hardware impairments. By feeding VFDT-derived I and Q trajectories into a CNN, the approach yields device fingerprints that generalize across location and channel variations, demonstrated on a real testbed of 30 Pycom devices using LoRa and WiFi. The results show substantial robustness to domain shifts and scalable performance as the number of devices increases, outperforming conventional IQ-based inputs in cross-domain settings. The method promises practical benefits for secure device identification and authentication in heterogeneous wireless environments, with potential extensions to larger-scale deployments and impairment disentanglement.

Abstract

RF data-driven device fingerprinting through the use of deep learning has recently surfaced as a possible method for enabling secure device identification and authentication. Traditional approaches are commonly susceptible to the domain adaptation problem where a model trained on data collected under one domain performs badly when tested on data collected under a different domain. Some examples of a domain change include varying the location or environment of the device and varying the time or day of the data collection. In this work, we propose using multifractal analysis and the variance fractal dimension trajectory (VFDT) as a data representation input to the deep neural network to extract device fingerprints that are domain generalizable. We analyze the effectiveness of the proposed VFDT representation in detecting device-specific signatures from hardware-impaired IQ (in-phase and quadrature) signals, and we evaluate its robustness in real-world settings, using an experimental testbed of 30 WiFi-enabled Pycom devices. Our experimental results show that the proposed VFDT representation improves the scalability, robustness and generalizability of the deep learning models significantly compared to when using IQ data samples.

Figures (14)

• Figure 1: CNN model classifier trained on data collected from 30 devices placed 1m away from the receiver, but tested on data collected when the devices are placed at different distances from the receiver: 1m, 2m, 3m, and two random locations. Experiments are taken indoor, in a lab environment.
• Figure 2: VFDT of the I signal component of Pycom devices.
• Figure 3: Basic RF transmitter front end.
• Figure 4: Impact of PA Nonlinearity Distortion on the VFDT of the IQ signals under varying IIP3 values.
• Figure 5: Impact of IQ Imbalance on the VFDT of the IQ signals under varying IQ amplitude imbalance values.