Enhanced EVD based channel estimation and pilot decontamination for Massive MIMO networks

The enhanced eigenvalue decomposition(EEVD) based channel estimation algorithm, which could solve the pilot contamination problem in massive multiple-input and multiple-output(Massive MIMO) channel estimation when the number of antennas at base stations(NABT) tends to infinity, is proposed in this paper. The algorithm is based on the close relationship between covariance matrix of received pilot signal and the channel fast fading coefficient matrix, i.e. the latter is the eigenvector matrix of the former when NABT tends to infinity. Therefore, we can get a set of normalized base vectors from the eigenvalue decomposition(EVD) of sample covariance matrix in practical Massive MIMO networks. By multiplying the received pilot signal with conjugate transpose of normalized base vector matrix, the channel matrix is projected to a lower dimensional matrix, and the intra-cell and inter-cell interference can be eliminated completely when NABT tends to infinity. Thus, we only need to estimate the lower dimensional projected matrix during the channel estimation. Simulation results show that the mean square error(MSE) performance of channel estimation is improved with approximately two orders of magnitude when the signal-to-noise ratio(SNR) is 40 d B, compared with EVD based channel estimation algorithm. And the signal-to-interference ratio(SIR) is improved greatly as well. The increment of SIR becomes larger and larger as SNR increasing.