Adversarial AI-Driven Covert Communications in Bistatic Backscatter Systems: An Overview

  • Tran, Hao Hoang; 
  • Do, Quang Tuan; 
  • Do, Tung Son; 
  • Oh, Junsuk; 
  • Cho, Sungrae
Citations

WEB OF SCIENCE

0
Citations

SCOPUS

0

초록

Bistatic backscatter communication is a promising technology for low-power Internet of Things (IoT) applications. However, its broadcast nature poses significant security risks. Covert communication aims to mitigate these risks by concealing the very existence of a transmission. This paper provides an overview of the evolution of covert communication schemes in bistatic backscatter systems, from classical analytical models to modern adversarial AI frameworks. We first review the foundational approach, where system parameters are optimized against a simple, model-based warden. We then detail the paradigm shift required to counter an intelligent warden that employs machine learning (ML) for detection. We propose a framework based on Generative Adversarial Networks (GANs) and Reinforcement Learning (RL), where a legitimate RL agent learns to generate signal characteristics that are indistinguishable to an advanced, data-driven warden. By surveying key literature, we compare these approaches, highlighting the necessity of AI-driven strategies for ensuring robust covertness in future wireless networks.

키워드

Bistatic Backscatter; Covert Communication; Generative Adversarial Network (GAN); Physical Layer Security; Reinforcement Learning
제목
Adversarial AI-Driven Covert Communications in Bistatic Backscatter Systems: An Overview
저자
Tran, Hao Hoang; Do, Quang Tuan; Do, Tung Son; Oh, Junsuk; Cho, Sungrae
DOI
10.1109/ICOIN68469.2026.11480488
발행일
2026
유형
Proceedings Paper
저널명
International Conference on Information Networking
페이지
226 ~ 231