Fast algorithm for estimating singular values of Hermitian matrix

Singular Value Decomposition (SVD) is often used in linear algebra and signal processing. SVD allows to decompose the original matrix into a product of three matrices, two of which are singular vectors, and one is a diagonal matrix of singular values. SVD allows to perform a denoising operation, isolate the main components of the signal, etc. There are many ways to calculate the singular values of a Hermitian matrix. The problem with these algorithms is their high computational complexity. This article will present a novel method that allows to calculate all singular values with low computational cost by using the knowledge of the first two singular values and the matrix trace. Efficiency of the proposed algorithm was evaluated in application to MMSE/IRC MIMO receiver approach