Research On Channel Estimation In Millimeter Wave Massive MIMO System

With the development of the fifth-generation mobile communication network,massive multiple input multiple output technology and millimeter wave technology are widely recognized.The channel contains state information during signal transmission,but it will be affected by noise and interference during transmission.The accuracy of channel estimation technology is related to the accuracy of signal recovery at the receiving end,so the research on channel estimation technology is of great significance.This paper studies the channel estimation algorithm in millimeter wave massive MIMO system.Through the study of deep learning and channel estimation algorithms,we can deeply understand the characteristics of different neural networks,and use their characteristics to combine with channel estimation algorithms to obtain better channel estimation algorithms.This paper proposes a channel estimation algorithm based on GAN-CNN Blind Denoising(GCBD)neural network.The unknown of noise is expanded by generating an adversarial neural network,and then denoising is performed by a convolutional denoising neural network to improve the performance of channel estimation.The channel estimation algorithm based on GCBD proposed through analysis and simulation can improve the channel estimation performance and can deal with different mixed noises.The Fully Convolutional Denoising Approximate Message Passing(FCDAMP)channel estimation algorithm proposed in this paper innovatively combines a fully connected convolutional denoising neural network with an approximate message passing algorithm based on learning.The channel noise level estimation map and the received signal containing noise are connected and input into the denoising subnet to achieve efficient denoising,thereby obtaining higher channel estimation accuracy.In the time varying channel system,the three-dimensional structure of the delay Doppler angle channel is obtained through orthogonal time-frequency space modulation technology.This paper proposes a 3-Dimensional Compressive Sampling Matching Pursuit(3D-Co Sa MP)channel Estimate algorithm.Through compressed sensing technology,the delay dimension,Doppler dimension and angle dimension are recovered respectively to obtain accurate channel estimation values.And verify the algorithm in terms of signal-to-noise ratio,different user speeds,number of base station antennas and pilot overhead rate.