Deterministic compressed sensing based channel estimation for MIMO OFDM systems

In most of the existing compressed sensing (CS) based channel estimation schemes for multiple-input multiple-output (MIMO) orthogonal frequency division multiplexing (OFDM) systems, the randomly allocated pilot is difficult to be implemented in real applications and introduces additional pilot overhead for transmitting the information on pilot locations which is required in channel reconstruction at receiver. In this paper, a channel estimation scheme based on deterministic compressed sensing is proposed to cut down the pilot overhead in MIMO OFDM systems. To be specific, a deterministic pilot placement scheme is proposed to select the subset of the subcarriers for pilot transmission. Since this deterministic pilot placement leads a new kind of deterministic measurement matrices in CS model, the mutual coherence property of the deterministic matrix is verified to establish theoretical guarantee for the pilot placement scheme. Then an improved reconstruction algorithm is proposed to match the structure of the deterministic matrix. Numerical results demonstrate that even without the pilot locations information, the proposed channel estimation scheme based on deterministic compressed sensing achieves similar estimation accuracy as conventional estimator with random pilot placement.