A Source Localization Method Using Complex Variational Mode Decomposition

Area of Science:

• Array Signal Processing
• Computational Electromagnetics
• Statistical Signal Processing

Background:

• Source localization using passive sensor arrays is crucial in various fields.
• Compressive sensing (CS)-based methods offer high resolution but require dense grids, leading to computational inefficiency.
• Existing CS methods struggle with coherent sources and limited snapshots.

Purpose of the Study:

• To develop a more efficient and accurate source localization technique for passive sensor arrays.
• To address the computational burden and limitations of traditional CS-based methods.
• To enable high-resolution localization of near-field, far-field, and mixed sources, even with coherent signals and single snapshots.

Main Methods:

• Extension of complex variational mode decomposition (CVMD) from nonstationary signal analysis to array signal processing.
• Modeling the source localization problem as a time-domain frequency-modulated signal.
• Decomposition of array measurements using CVMD to isolate signals from different source locations.
• Model fitting to decomposed subsignals for estimating source direction and range.

Main Results:

• The proposed CVMD-based method successfully localizes pure far-field, pure near-field, and mixed near-field/far-field sources.
• Achieves high-resolution localization for coherent sources using only a single snapshot.
• Demonstrates significantly reduced computing time compared to conventional CS methods.
• Maintains accuracy and robustness under challenging conditions.

Conclusions:

• CVMD offers a computationally efficient and high-resolution alternative for passive sensor array source localization.
• The method effectively handles complex scenarios including coherent sources and limited data.
• This approach advances the capabilities of array signal processing for accurate source identification.