A Total Lp-Norm Optimization for Bearing-Only Source Localization in Impulsive Noise with SαS Distribution

Area of Science:

• Signal Processing
• Acoustics
• Estimation Theory

Background:

• Bearing-only source localization is challenging in impulsive noise.
• Existing Iteratively Reweighted Pseudolinear Least-Squares (IRPLS) methods suffer from bias due to system matrix and noise vector correlation.

Purpose of the Study:

• To propose a novel Total Lp-norm Optimization (TLPO) method for robust bearing-only source localization.
• To mitigate bias in source localization caused by impulsive noise with symmetric α-stable distribution.

Main Methods:

• Developed the Total Lp-norm Optimization (TLPO) method by minimizing errors in system matrix and data vector.
• Derived an equivalent form of TLPO.
• Implemented two algorithms: Iterative Generalized Eigenvalue Decomposition (IGED) and Generalized Lagrange Multiplier (GLM) to solve the TLPO problem.

Main Results:

• The proposed TLPO method effectively reduces bias in bearing-only source localization.
• Numerical examples show IGED and GLM algorithms outperform the traditional IRPLS algorithm.
• Demonstrated improved performance in impulsive noise environments with symmetric α-stable distributions.

Conclusions:

• The TLPO method offers a more robust and accurate approach to bearing-only source localization.
• IGED and GLM algorithms provide efficient solutions for the TLPO problem.
• This research contributes to improved underwater acoustic localization and tracking systems.