Massive MIMO Uplink Channel Estimation using Compressive Sensing

Abstract

Over the past decades, wireless communication systems have made a real revolution. Huge amounts of information circulate on the networks every second. To meet the growing demand for data throughput, massive multiple-input multiple-output (MIMO) has been proposed as a key technology for future wireless communications by offering substantial energy gain and spectral capacity. However, the advantages of such a system generally requires that the channel state information should be available on the transmitter (CSIT). For this purpose, this work presents the implementation of a new channel estimation technique based on training sequence in Time Division Duplex (TDD) mode. As it is known, wireless communication tends to be sparse, that is why, in this paper, we will focus on this property and propose a new approach based on compressed sensing technique to estimate the Channel Impulse response (CIR) of the massive MIMO system. As a simulation results, it has been shown that the proposed algorithm Block Orthogonal Matching Pursuit (B-OMP) outperforms the old approach known as Adaptive Orthogonal Matching Pursuit (AOMP) especially in terms of Bit Error Rate (BER) where it makes a gain of 1 dB over AOMP algorithm. However, in terms of Normalised Mean Square Error (NMSE), both of them have shown the same performance, where both of curves are superimposed. On the other hand and in order to underline the performance of the proposed approach, a comparison in terms of computational complexity of the two algorithms has been done and it has proved that the complexity of B-OMP is in the order of only O(KN p N t L) while the complexity of AOMP is equal to O(KN p N t UL) so the AOMP requires N t more iteration to recover the estimated signal.