Wide-band signal parameter estimation based on higher-order statistics

In this paper we develop an algorithm to improve the accuracy of the wideband signal parameter estimation. It is well known that in the presence of an unknown noise, these estimates may be grossly inaccurate. The proposed algorithm uses both the fourth order cumulant for the suppression of the gaussian noise, the transformation matrices for estimating the coherent cumulant matrix and a noneigenvector algorithm for the characterization of the sources. We show that the performances of bearing estimation algorithm improve substantially when the proposed algorithm is used. This method is tested on simulated data and its performances are clearly pointed out.