Iterative sparse channel estimator based on SpaRSA approach

In this paper, an iterative sparse channel estimation for orthogonal frequency division multiplex (OFDM) communication system is investigated based on the sparse reconstruction by separable approximation (SpaRSA), which is regarded as one of the fastest algorithms for l2-lj problem and can obtain its global optimal solution. The proposed estimator comprised of thresholding is applied to detect channel taps. Then, a modified SpaRSA with adaptive regularization parameter is used to refine the estimation of nonzero channel taps. Simulation results for typical sparse channels show effectiveness of the proposed algorithm over other existing methods.