Research On Subspace-based Channel Estimation Of MIMO-OFDM System

As the focus of research on the next generation mobile communication, MIMO-OFDM technology can greatly improve the communication quality of the wireless channel by employing time, frequency and spatial diversity techniques. The technique of channel estimation is a forecast method for time- varying wireless channel state information(CSI), which provide supporting evidence for the subsequent signal detection, channel equalization and spatial-time decoding.In order to get a more accurate estimation result,a small amount of received symbol block subspace estimation method is referenced at the expense of a little more algorithms complexity. The proposed algorithms is a novel extension to the noise subspace method based on repetition index by adding additional columns in the matrix equation used for subspace decomposition. Simulation results show that the proposed algorithm exhibits NRMSE decreases 60.1%,and the number of received symbol blocks decreases from 96 to 20 compared with those of conventional subspace method with large receive data blocks in a MIMO-OFDM system with two transmit antennas and three receive antennas respectively. The proposed algorithm can not only improve the channel estimation performance, but also decrease the required number of received OFDM symbols.Besides, by studying the noise subspace method based on repetition index, the MIMO-OFDM based repetition index method with a new signal permutation method is proposed. which first restructure autocorrelation matrix with few data blocks by explo iting a new permutation of the received signal, and then obtain blind channel response according to orthogonal property of signal subspace and noise subspace, estimate the ambiguity matrix by using LS method finally. Simulations results show that, compared with the conventional noise subspace method which needs 48 received symbol blocks, the value of the normalized root mean square error has decreased 43.2% and 58.4%, while the receiving blocks is 12 and 20 in the new improved algorithm respectively.