SPARSE SIGNAL RECONSTRUCTION USING REFINED INSTANTANEOUS FREQUENCY ESTIMATION

This study proposes an instantaneous frequency (IF) estimation and then applies this algorithm to reconstruct missing samples. The proposed IF estimation and sparse reconstruction algorithm is developed by adding a refinement step where the crude estimates of IFs and signal components are further improved through the re-estimation stage. During the re-estimation stage, both IFs and the signal components are re-estimated by removing all the remaining components. Experimental results indicate that the proposed strategy improves the accuracy of IF estimates and the reconstruction process of signals with missing samples.