Direction-of-Arrival Estimation via Sparse Bayesian Learning Exploiting Hierarchical Priors with Low Complexity.

Area of Science:

• Signal Processing
• Array Signal Processing
• Computational Electromagnetics

Background:

• Sparse Bayesian Learning (SBL) excels in direction-of-arrival (DOA) estimation but traditional Gaussian priors limit sparse signal recovery (SSR).
• Hierarchical priors improve SSR but struggle with multiple measurement vector (MMV) data.
• Existing methods face challenges with MMV data and computational complexity.

Purpose of the Study:

• To develop a novel block-sparse SBL (BSBL) method for DOA estimation in MMV scenarios.
• To enhance SSR performance by combining hierarchical priors with a block-sparse model.
• To reduce the computational complexity of BSBL for practical applications.

Main Methods:

• A block-sparse SBL (BSBL) method is proposed, vectorizing the MMV model into a block-sparse form.
• Hierarchical priors are integrated to improve SSR capabilities.
• Two low-complexity variants, BSBL-APPR (approximation) and BSBL-GAMP (generalized approximate message passing), are introduced.

Main Results:

• BSBL demonstrates superior SSR performance compared to traditional SBL and other state-of-the-art algorithms in MMV models.
• BSBL-APPR and BSBL-GAMP significantly reduce computational complexity while maintaining high estimation accuracy.
• The proposed methods effectively suppress temporal correlation and handle wideband sources.

Conclusions:

• The developed BSBL method offers enhanced DOA estimation performance for MMV data by leveraging hierarchical priors and a block-sparse structure.
• BSBL-APPR and BSBL-GAMP provide efficient and computationally feasible solutions for complex sparse signal recovery problems.
• The BSBL framework shows promise for advanced applications in array signal processing and beyond.