Multi resolution signal analysis using improved Wigner Ville Distribution

Abstract

This paper presents an idea for improving the multi resolution performance of Wigner Ville Distribution during multi component signal analysis. The method concentrates on reducing the mathematical cross-terms of a multi-component signal, which are unfortunate terms arised in the calculation of Wigner Ville Distribution (WVD). Linear FM signals without any noise component ideally have a clear line representation while plotted in the time-frequency domain. When the WVD of a multi-component signal is calculated, the resultant cross-terms will sometimes have higher peaks compared to the main signal components. There have been many methods proposed to filter out these cross terms, so that the signal can be clearly represented in the time-frequency domain. In this paper, a method based on Hough Transformation is applied to WVD of the signal. A line in the time-frequency domain can be represented as a point in the Hough transformed domain. Once we identify the coordinates of this point in the transformed domain, we can estimate the properties of the line in the time-frequency domain which corresponds exactly to the actual signal components. This paper proposes the back parameter estimation method to select and retain only the components corresponding to the auto-terms of the signal, from the time-frequency domain of WVD.