Channel estimation for MIMO-OFDM systems based on Subspace Pursuit algorithm

Abstract

MIMO-OFDM technique combines the advantages of MIMO and OFDM and is widely used in high data rate system. Conventional linear channel estimators are considered optimal under the assumption of rich multipath, while the practical physical multipath channels tend to exhibit sparse structure. By exploiting the coherent sparsity of the multipath channels, channel estimation based on compressive sensing (CS) method can greatly decrease the pilot overhead burden. In this paper, we present CS-based channel estimation method by exploiting the time-domain sparsity and one of the CS algorithms-Subspace Pursuit (SP) in order to estimate the channel impulse response in MIMO-OFDM system. Simulation results demonstrate a significant reduction of the number of pilots compared with least squares (LS) channel estimation.