Sparse Bayesian learning for DOA estimation with mutual coupling.

Area of Science:

• Signal Processing
• Array Signal Processing
• Statistical Inference

Background:

• Sparse Bayesian Learning (SBL) is crucial for direction-of-arrival (DOA) estimation.
• The assumption of a precisely known measurement matrix in SBL is often violated in practice due to mutual coupling.
• Existing methods struggle with unknown or misspecified mutual coupling, impacting DOA estimation accuracy.

Purpose of the Study:

• To develop a modified SBL method for joint estimation of DOAs and mutual coupling coefficients.
• To address the challenge of imperfect array manifolds in uniform linear arrays (ULAs).
• To improve the robustness and accuracy of DOA estimation in the presence of mutual coupling.

Main Methods:

• A modified Sparse Bayesian Learning (SBL) approach is proposed.
• Utilizes a hierarchical Student t prior to enforce signal sparsity.
• Employs a distinct Bayesian inference for the Expectation-Maximization (EM) algorithm to update mutual coupling coefficients.
• Incorporates Singular Value Decomposition (SVD) for computational efficiency and noise reduction.

Main Results:

• The proposed method enables joint estimation of DOAs and mutual coupling coefficients.
• The hierarchical Student t prior enhances sparsity enforcement.
• The EM algorithm provides efficient updates for mutual coupling coefficients.
• SVD integration reduces computational complexity and improves noise sensitivity.

Conclusions:

• The modified SBL method effectively handles imperfect array manifolds caused by mutual coupling.
• This approach offers a more robust and computationally efficient solution for DOA estimation with ULAs.
• The joint estimation of DOAs and mutual coupling coefficients improves overall system performance.