Frequency-domain estimation of time-domain correlation matrix for MIMO-OFDM systems

Since the received MIMO-OFDM signal is usually corrupted in the time-domain due to some adverse factors such as frequency offset and large peak-to-average power ratio (PAPR) etc., a high-quality signal may only be obtained in the frequency domain. However, the second-order statistics of the time-domain signal are very often used in a blind or semi-blind channel estimation, which means that an IFFT processor is required in the receiver to achieve the time-domain signal. This additional IFFT incurs a high computational complexity and a long time delay in real-time communication systems. In this paper, we propose a new algorithm for the computation of the time-domain correlation matrix directly from the received frequency-domain signal. The proposed algorithm is proven to be equivalent to the original algorithm of estimating the time-domain correlation matrix, while the former can avoid an IFFT process when high-quality signal is only available in the frequency domain.