A Novel Fast High Resolution Music Algorithm

Abstract

The paper describes new techniques to determine the number of sources for a signal based on LU and QR decomposition. We propose novel methods to calculate the threshold for noise subspace estimation used in high resolution array processing methods without eigenvalue decomposition. The paper states that previous techniques primarily use eigenvectors and eigenvalues. We propose an approximation of MUSIC algorithm. This approximation decreases the computational complexity. A full mathematical evaluation of the technique is provided and simulations show that the approach is effective.