Skip to main navigation Skip to search Skip to main content

Deception Against Reactive Jammer: Deep Reinforcement Learning for Adaptive Anti-Jamming

  • Xintai Cao
  • , Qubeijian Wang
  • , Xiuping Li
  • , Wen Sun
  • , Yalin Liu

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

Reactive jamming is a stealthy, highly disruptive threat in wireless networks that evades simple detection by remaining dormant until legitimate transmissions occur. Deception-based anti-jamming strategies can induce premature jammer activation by transmitting deceptive signals. However, existing work has largely overlooked the design of the transmitting window of deception signals, resulting in the waste of channel resources or the failure of deception. To address this gap, we propose a dynamic deception-based anti-jamming scheme, modeling the interaction as a two-phase Markov Decision Process (MDP). Our approach leverages a Deep Q-Network (DQN) to intelligently and adaptively determine the optimal deception window length, then observes the reaction from the jammer to dynamically select the transmission mode, maximizing channel capacity. Experimental results demonstrate that, compared with fixed-window strategies, the proposed DQN-based scheme achieves significantly improved anti-jamming performance.

Original languageEnglish
Title of host publicationICC 2026 - IEEE International Conference on Communications, Proceedings
ISBN (Electronic)9798319542090
DOIs
Publication statusPublished - 2026
Event2026 IEEE International Conference on Communications, ICC 2026 - Glasgow, United Kingdom
Duration: 24 May 2026 → 28 May 2026

Publication series

NameIEEE International Conference on Communications
ISSN (Print)1550-3607

Conference

Conference2026 IEEE International Conference on Communications, ICC 2026
Country/TerritoryUnited Kingdom
CityGlasgow
Period24/05/26 → 28/05/26

Keywords

  • DRL
  • dynamic deception
  • reactive jamming

Fingerprint

Dive into the research topics of 'Deception Against Reactive Jammer: Deep Reinforcement Learning for Adaptive Anti-Jamming'. Together they form a unique fingerprint.

Cite this