Polynomial coefficient finding for Root-MUSIC

Abstract

Root-MUSIC (MUltiple SIgnal Classification) is the polynomial rooting form of MUSIC, namely, the spectrum peak searching is resplaced by the polynomial rooting in MUSIC implementation. The coefficients finding of the polynomial is the critical problem for Root-MUSIC and its improvements. By analyzing the Root-MUSIC algorithm thoughly, the finding method of the polynomial coefficient is deduced and the concrete calculation formula is given, so that the speed of polynomial finding roots will get the bigger exaltation. The particular simulations are given and attest correctness of the theory analysis and also indicate that the proposed algorithm has preferable estimating performance.

Grid‐less coherent DOA estimation based on fourth‐order cumulants with Gaussian coloured noise

“…But it has restrictions to practical applications due to the high computational cost. Root methods [3–5] construct polynomials and then estimate DOAs through polynomial rooting, which exhibit high resolution capability. The process of peak searching can be avoided by using the polynomial parameterisation.…”

“…The process of peak searching can be avoided by using the polynomial parameterisation. The most commonly used method is root‐MUSIC [3]. He et al [4] introduce a procedure of locating polynomial roots based on the Vandermonde characteristic of the uniform linear array (ULA) manifold.…”