Decomposition of Time-varying Multicomponent Signals using Time-Frequency Based Method

This paper proposes a new time-frequency based signal decomposition method. This approach is based upon the eigenvalue decomposition method, combined with the s-method that produces a sum of the Wigner distribution of individual signal components. The proposed decomposition method is theoretically derived. The efficiency and accuracy of the proposed decomposition method is demonstrated on experimental high-frequency surface-wave radar data sets. The method presented here is not restricted to this application, but it can also be applied in various other settings of non-stationary signal analysis and filtering, for example, moving target identification in synthetic aperture radar (SAR) and inverse synthetic aperture radar (ISAR) images, etc