Improved Time-Domain Threshold Determination for Sparse Channel Estimation in OFDM System

Channel estimation is one of the most essential requirements in Orthogonal Frequency Division Multiplexing (OFDM) system and its exactness severely impacts the performance of the entire system. The most significant tap (MST) selection approach requires prior knowledge of channel statistics which increases complexity and a reduced spectral efficiency. Therefore, in order to tackle the approach of sparse channel estimation in OFDM systems, an improved time domain threshold (ITDT) for sparse channel estimation is thus proposed. The proposed ITDT aims at achieving an improved channel estimation performance with no prior knowledge of channel state information and noise standard deviation. At the start, an initial channel impulse response is obtained by Least Square (LS) approach. Next, a novel ITDT is proposed to obtain the estimated coefficient of noise by extracting the non-zero channel taps. Simulation results indicate a better normalized mean square error (NMSE) of 20-dB gain for a same NMSE for the proposed ITDT compare to the non-sparse LS time domain estimator.