A Deep Learning Detector for 3D Massive MIMO Systems with Impulsive Noise

In this paper, the signal detection problem of the Three dimensional Massive Multiple-input multiple-output (3D Massive MIMO) wireless communication system with impulsive noise is studied. In most signal detection methods, the assumed noise model is AWGN, but in the actual scene, the noise is not ideal, and is probably impulsive noise. Therefore, in the actual signal detection, the performance of the traditional detection mechanism may decline due to the change of noise model. In order to solve this problem, we propose a signal detection network based on the correntropy function, which can overcome the influence of impulsive noise. In addition, the size of the channel state information matrix (CSI) significantly increases the computational complexity. We use the sparse characteristics of the angular domain 3D massive MIMO channel to compress CSI, which reduces the complexity of the system without obviously affecting the performance.