Deep Reinforcement Learning-Based Intelligent Reflecting Surface for Cooperative Jamming Model Design
Shaofang Lu,Xianhao Shen,Panfeng Zhang,Zhen Wu,Yi Chen,Li Wang,Xiaolan Xie,ShaoFang Lu
DOI: https://doi.org/10.1109/access.2023.3312546
IF: 3.9
2023-09-15
IEEE Access
Abstract:Owing to the nature of wireless channels, wireless transmission is vulnerable to attacks by adversaries; therefore, security has always been a critical issue in wireless networks. In this context, intelligent reflecting surfaces (IRS), as an emerging and promising technology, synergize with physical layer security (PLS), offering novel avenues to enhance privacy and resistance against interference in wireless communication. This paper investigates a cooperative jamming communication model assisted by IRS. Under the constraints of minimum safe rate and inaccurate channel state information (CSI), a deep reinforcement learning (DRL)-based framework is proposed to jointly optimize the BS transmitting beamforming power distribution and IRS phase shift matrix to maximize the system energy efficiency. We first formulate an anti-jamming communication optimization problem as a Markov decision process (MDP) framework and then design a DRL-based algorithm, in which the joint design is obtained through trial-and-error interactions with the environment by observing predefined rewards in the context of continuous state and action to generate an optimal policy. The simulation results show that when the number of IRS components is increased from 20 to 100, the proposed scheme can improve energy efficiency by 40.1%, which is better than other schemes.
computer science, information systems,telecommunications,engineering, electrical & electronic