Enhanced Channel Estimation for MIMO OFDM Systems

Abstract

In this paper, the enhanced approximate linear minimum mean square error channel estimation (EALMMSE) method is proposed. It outperforms the conventional approximate linear minimum mean square error channel estimation (ALMMSE) method, known for its poor performance for the low signal-to-noise ratio (SNR) regime. The proposed EALMMSE method uses the channel autocorrelation matrix in the frequency domain, which estimates channel coefficients more accurately by the Pilot-Only Autocorrelation Matrix. The EALMMSE achieves improved performance for a broad range of SNR, offsetting the shortcoming of the conventional ALMMSE method so that it can be applied to various practical network environments. In addition, the EALMMSE method is well suited for a multiple-input-multiple-output (MIMO) OFDM system. It is demonstrated by simulation that EALMMSE outperforms ALMMSE under the standard channel model developed by the Third-Generation Partnership Project (3GPP).