基于深度强化学习的多RIS辅助干扰信道传输方法

Transmit Design for Multi-RIS-Assisted Inference Channel:A DRL Approach

无线覆盖扩展是6G的核心任务。RIS作为新兴智能无线环境技术,可以有效缓解障碍物对高频信号传播的影响。然而,RIS通信系统也面临信道估计困难和资源优化复杂的挑战。为此,研究面向多RIS辅助干扰信道系统的能量效率最大化问题。为降低信道估计频次,考虑信道统计信息;为求解所提出问题,采用深度强化学习方法。将所考虑问题建模为MDP,并提出混合最大合并比和破零方法设计动作空间;在此基础上,基于近端策略优化算法训练智能体求解MDP。实验结果展示了所提出方法在不同信噪比下的收敛情况与适应能力。

Enhancing wireless coverage is the core task of the sixth-generation mobile communication system(6G).Reconfigurable intelligent surface(RIS),as an emerging smart radio environment enabler,is capability of effectively alleviating the impact of obstacles on the high-frequency signal propagation.However,RIS communication systems also face challenges in channel estimation and complex resource optimization.Therefore,this paper studies the energy efficiency maximization problem for multi-RIS-assisted interference channel systems.To avoid frequent channel estimation,the channel statistical information for RIS-user links is considered.To solve the proposed problem,a deep reinforcement learning method is adopted.The problem is modeled as a Markov decision process(MDP)and a hybrid method of the maximum ratio combining and zero forcing is proposed to design the action space.Then,the proximal policy optimization algorithm is used to train the agent to solve the MDP.Experimental results demonstrate the convergence and adaptability of the proposed method under different signal-to-noise ratios.