Signal denoising using Hankel matrix rank reduction

In this paper, signal denoising is achieved through a procedure consisting of rank reduction of the signal's Hankel matrix, then Hankel matrix restoration and finally projecting the enhanced signal on an optimally chosen signal space to get rid of the orthogonal noise components. The process can be iterated until the changes in the singular values of the Hankel matrix are less than a specified threshold. Illustrative examples are given to verify that this process effectively reduces associated noise.