Underdetermined DOA Estimation Using MVDR-Weighted LASSO.

Area of Science:

• Signal Processing
• Array Signal Processing
• Compressive Sensing

Background:

• Direction of Arrival (DOA) estimation is crucial in various applications, including radar, sonar, and wireless communications.
• Traditional DOA methods often require prior knowledge of the number of sources or struggle with spatially-close or coherent sources.
• Compressive Sensing (CS) offers a promising framework for sparse signal recovery, applicable to DOA estimation.

Purpose of the Study:

• To formulate the DOA estimation problem within a compressive sensing framework.
• To introduce an extended array aperture to enhance the degrees of freedom.
• To propose a novel LASSO-based algorithm for DOA estimation without prior source number knowledge.

Main Methods:

• Formulation of DOA estimation in a compressive sensing (CS) framework.
• Application of Ordinary Least Square Adaptable Least Absolute Shrinkage and Selection Operator (OLS A-LASSO) for DOA estimation.
• Development and application of a new Minimum Variance Distortionless Response (MVDR) A-LASSO algorithm for DOA estimation in CS.

Main Results:

• The proposed MVDR A-LASSO algorithm effectively estimates DOA without relying on Singular Value Decomposition (SVD) or subspace orthogonality.
• The algorithm can estimate a high number of sources (up to ((M^2 - 2)/2 + M - 1)/2 with M sensors) without prior assumptions.
• Superior performance is demonstrated compared to classical DOA methods, particularly in low Signal-to-Noise Ratio (SNR), spatially-close, and coherent scenarios.

Conclusions:

• The novel MVDR A-LASSO algorithm provides a robust and flexible solution for DOA estimation within the CS framework.
• This method overcomes limitations of traditional techniques, enabling accurate DOA estimation even with unknown numbers of sources and challenging signal conditions.
• The enhanced array aperture and new LASSO algorithm significantly improve DOA estimation performance.