A Comprehensive Survey on Feature Extraction Techniques Using I/Q Imbalance in RFFI

TL;DR

This survey reviews feature extraction methods using I/Q imbalance for RF fingerprinting in IoT, highlighting challenges, recent advances, and ML/DL techniques to improve device authentication security.

Contribution

It provides a comprehensive overview of I/Q imbalance-based feature extraction techniques and discusses recent ML/DL approaches for enhancing RF fingerprinting in IoT.

Findings

I/Q imbalance can serve as a reliable hardware fingerprint.

Machine learning techniques improve identification accuracy.

Challenges include standardization and environmental variability.

Abstract

The proliferation of Internet of Things (IoT) devices has increased the need for secure authentication. While traditional encryption-based solutions can be robust, they often impose high computational and energy overhead on resource-limited IoT nodes. As an alternative, radio frequency fingerprint identification (RFFI) leverages hardware-induced imperfections-such as Inphase/Quadrature (I/Q) imbalance-in Radio Frequency (RF) front-end components as unique identifiers that are inherently difficult to clone or spoof. Despite recent advances, significant challenges remain in standardizing feature extraction methods, maintaining high accuracy across diverse environments, and efficiently handling large-scale IoT deployments. This paper addresses these gaps by providing a comprehensive review of feature extraction techniques that utilize I/Q imbalance for RFFI. We also discuss other hardware-based RF fingerprinting sources, including power amplifier nonlinearity and oscillator imperfections, and examine modern machine learning (ML) and deep learning (DL) approaches that enhance device identification performance.