Deep Learning for Intelligent Reflecting Surfaces Aided MIMO Systems

The intelligent reflector has recently attracted many attention and research enthusiasm because it can reduce the complexity of the hardware and the implementation cost compared to the traditional relay devices. The introduction of deep learning (DL) in intelligent reflection surface (IRS) is the current research hotspot and direction. In this paper, we establish an IRS system model based on known communication system knowledge. Then, a multi-layer perceptron neural network which consists of 8 hidden layers is involved to train the IRS system. Simulation results confirm the effectiveness of the proposed model training the IRS system.