OFDM-based Massive MIMO Channel Estimation using Gaussian Mixture Learning and Compressed Sensing Methods

Abstract

Massive MIMO-OFDM system is proved to be an effective and most sustainable technology to forthcoming applications of 5G wireless communications. It furnished significant gains that facilitate a higher number of user connections at high data rates with improved latency and reliability. To achieve accurate channel knowledge, lessen pilot overhead is necessary. To resolve this problem, one of the favorite approaches is compressed sensing. Sparse channel estimation develops the essential sparsity between the communicating channels that can be improved by the channel estimation efficacy with lower pilot overhead. To achieve this, non-zero vector distribution can be taking into consideration the Gaussian mixture accordingly, learn their characteristics towards the expectation-maximization procedure. The results of simulation have proved the performance of proposed estimation approach of channel keeping with minimum pilot overhead and developed exceptional symbol error rate (SER) performance of the system.