Structured Compressive Sensing Based Block-Sparse Channel Estimation for MIMO-OFDM Systems

In this paper, a compressive sensing based method named Priori-Information Aided Modified-SAMP algorithm is proposed to solve the problem of channel estimation in MIMO-OFDM systems. Firstly, coarse channel state information (CSI) as a priori-information of channel is obtained by using the complete pseudo-random noise (PN) sequences. Due to noise and the interference among antennas caused by the non-orthogonality of PN sequences, then, the accuracy of channel estimation is not so high that the priori-information aided modified-SAMP algorithm based on the obtained CSI is proposed to estimate CSI more accurately in temporal domain. Though the proposed method is based on the sparsity adaptive matching pursuit (SAMP) algorithm, there are some significant differences with each other in signal structure, support set selection, and adaptive step size etc. Theoretical analysis shows that the proposed algorithm has good convergence, moderate computational complexity and less training sequence overhead. Finally, the performance of the proposed method is verified through experimental simulations which show that compared with other algorithms, especially the orthogonal matching pursuit algorithm, the proposed algorithm not only improves the estimation accuracy but also greatly reduces the training sequence overhead.