Generative AI for Physical-Layer Authentication

Abstract

In recent years, Artificial Intelligence (AI)-driven Physical-Layer Authentication (PLA), which focuses on achieving endogenous security and intelligent identity authentication, has attracted considerable interest. When compared with Discriminative AI (DAI), Generative AI (GAI) offers several advantages, such as fingerprint data augmentation, fingerprint denoising and reconstruction, and protection against adversarial attacks. Inspired by these innovations, this paper provides a systematic exploration of GAI's integration into PLA frameworks. We commence with a review of representative authentication techniques, emphasizing PLA's inherent strengths. Following this, we revisit four typical GAI models and contrast the limitations of DAI with the potential of GAI in addressing PLA challenges, including insufficient fingerprint data, environment noises and inferences, perturbations in fingerprint data, and complex tasks. Specifically, we delve into providing GAI-enhance methods for PLA across the data, model, and application layers in detail. Moreover, we present a case study that combines fingerprint extrapolation, generative diffusion models, and cooperative nodes to illustrate the superiority of GAI in bolstering the reliability of PLA compared to DAI. Additionally, we outline potential future research directions for GAI-based PLA.

Tasks

Reproductions