MIMO-OFDM channel estimation based on time-domain window filter

Multiple transmit-and-receive antennas can be used in OFDM systems to improve communication quality and capacity. In this paper, by virtue of a linear matrix algebraic model, channel estimation for MIMO-OFDM can be treated as for SISO-OFDM. According to this model, we derive the least squares (LS) channel estimation algorithm assisted pilot sequences. Then, a novel time-domain windowing algorithm for antenna pair of MIMO is developed. In the time domain, through peak-detection processing, the possible peaks relative to multipath components are found and a proper window function is adopted to localize the peaks and minimize the noisy impact of non-peak area, and the filtered pilot signal is then interpolated and transformed back to the frequency domain. This algorithm discards the additive noise interference in the MIMO channel and approaches the real channel response without additive noise more closely, improving the SNR of the received signal after the deconvolution based on the channel estimation. The simulation results adopting this algorithm show the expected improvement of SNR of received signal.