A Novel Subspace Channel Estimation with Fast Convergence for ZP-OFDM Systems

Novel subspace channel estimation with fast convergence is proposed for single-input multiple-output (SIMO) zero-padded orthogonal frequency-division multiplexing (ZP-OFDM) systems. Stacking all the received OFDM symbols turns the channel matrix to be block Toeplitz. The block matrix scheme is then presented so a group of sub-vectors is formed from the stacked OFDM signal and the number of samples is increased. With the sub-vector samples, the perturbations of the sample correlation matrix and the noise subspace are reduced considerably and therefore, the channel estimation error is lowered. Computer simulations demonstrate that the proposed channel estimation is effective and robust compared with existing methods.