Subspace approach to blind channel estimation of MIMO-OFDM system

In order to overcome the shortcomings of traditional subspace algorithm that required lots of receiving symbols to compute the autocorrelation matrix and had slow convergence and so on, this paper presented a new fast convergence of subspace-based blind channel estimation algorithm for the Multi-Input Multi-Output Orthogonal Frequency Division Multiplexing (MIMO-OFDM) wireless communication system. Through the block matrix scheme, the channel matrix was turned to be block Toeplitz matrix. Without changing the block Toeplitz matrices, a large number of sub-vectors were formed from the stacked OFDM signal. By using the sub-vector samples, an accurate estimation of auto-correlation matrix could be obtained with fewer received block. The proposed method had the advantages of high estimation accuracy and low computational complexity inherited from traditional subspace algorithm, but also relaxed the channel time-invariant requirement and accelerated the convergence speed. Theoretical analysis and simulation results show the effectiveness and better bit error rate performance of the proposed method．