Deep SIC Separator Based Beamforming in MmWave Massive MIMO

Abstract

Millimeter wave has become popular due to its unique sparse characteristics and abundant frequency band resources. Hybrid precoding has been used in millimeter wave (mmWave) massive multiple-input multiple-output (MIMO) system to reduce the high cost and power consumption of full digital precoding. However, conventional hybrid precoding requires complex channel estimation before beamforming, which needs pretty high pilot consumption or complex computation. This paper combines Successive Interference Cancellation(SIC)and deep learning, and proposes a deep learning structure based on the idea of successive interference cancellation. This method does not require perfect channel state information, thus greatly reducing the cost of pilots. The simulation results show that when the number of antennas is large, this method can maintain a fixed low complexity, which is better than other methods.