A robust channel estimator at the high Doppler frequency via matching pursuit technique

In this paper, we propose a robust channel estimator for fast time-varying channels. For this technique, a basis expansion model (BEM) is used in order to approximate the time-variant channel. Pilot symbol assisted modulation (PSAM) is employed as the training scheme, and the coefficients of the BEM are obtained via the matching pursuit (MP) algorithm. Through computer simulations, it is demonstrated that MP is able to offer a reliable channel estimate even when it selects an inappropriate number of basis functions, whilst it was previously reported that more basis functions than necessary could cause performance degradation by enhancing the noise. Thus, MP is robust with respect to mismatch in the number of basis functions in the BEM. Furthermore, the simulation results show that MP can operate well even without knowledge of the Signal-to-Noise Ratio (SNR).