A Novel Sparse Channel Estimation Method Based on Discriminant Analysis for OFDM System

Owning to the advantage of an increase in spectral efficiency by reducing the transmitted pilot tones, compressed sensing has been widely applied to pilot-aided sparse channel estimation in OFDM systems. In this paper, we focus on increasing the accuracy of channel estimation and propose a novel sparse channel estimation method that takes the effect of the additive noise into consideration. In the proposed method, the channel impulse response (CIR) is represented as a combination of ideal channel paths and additive noise. We firstly use the Orthogonal Matching Pursuit (OMP) algorithm to estimate the path delays and path gains. Then, the Mahalanobis distance discriminant analysis is applied to channel paths identification. The proposed method could effectively distinguish the significant channel paths especially those with small amplitude from the noise. Simulation results demonstrate the effectiveness of the proposed sparse channel estimation method. Compared with the conventional CS-based channel estimation algorithms, the new method proposed here enjoys superior performance in terms of bit error rate (BER) and mean square error (MSE).