MIMO-OFDM channel estimation based on minimum error entropy criterion under non-Gaussian environment

Abstract

ABSTRACT Using the orthogonal frequency division multiplexing (OFDM) technique, a multiple-input multiple-output (MIMO) system can provide high spectral efficiency and high data transmission over a fading channel. To achieve the best performance in a MIMO-OFDM system, high accuracy in channel estimation is a very important factor which leads to appropriate receiver design. Therefore, in the most channel estimation algorithms, the mean square error (MSE) is the main criterion for noise minimisation which is robust in the case of Gaussian noise. However, in telecommunication systems which do not have noise with Gaussian distribution, the MSE criterion is not appropriate. So as to tackle this problem, a robust adaptive filtering algorithm was proposed using minimum error entropy (MEE) criterion and improved least square (ILS) approach which by far is better than MMSE criterion to robust channel estimation. In this work, impulsive noise is used for non-Gaussian environment simulation. Furthermore, MEE is employed to attenuate the impulsive noise, and ILS is adopted to reduce LS estimation variance. Compared to the MSE-based algorithms, simulation results indicate that the proposed method outperforms the channel estimation in a non-Gaussian environment.