Cross-terms Reduction in WVD using Sliding Window-based Variational Mode Decomposition

Abstract

Wigner-Ville distribution (WVD) has become one of the powerful tools in the time-frequency analysis of non-stationary signals. Still, the presence of cross-terms in WVD for multi-component signals limits its applicability and interpretation. Variational mode decomposition (VMD) has been applied to remove the cross-terms from WVD. However, it fails to remove cross-terms between the spectral components from the same frequency band but with different time signatures. This paper proposes a novel sliding window-based VMD (SW-VMD) technique for removing cross-terms from WVD by overcoming the drawbacks of the VMD-based WVD method. The proposed method segments a multi-component signal using overlapping windows and each windowed signal is decomposed into intrinsic mode functions using VMD. The WVDs of all the IMFs are added together to obtain the cross-terms free WVD. Energy scaling is also performed to minimize the effect of overlapping windows. The performance of the proposed method is assessed for different multi-component synthetic signals and real-world ECG signals using different performance measures. The simulation results reveal that the SW-VMD-based WVD method can effectively remove the cross-terms from the WVD with a better autoterms information and time-frequency resolution. Results from the proposed method are also compared with the VMD-based WVD method to exhibit supremacy of the proposed method.