On the Error Metrics Used for Direction of Arrival Estimation.

Area of Science:

• Signal Processing
• Array Signal Processing
• Acoustic and Electromagnetic Wave Propagation

Background:

• Evaluating Direction of Arrival (DoA) estimation performance relies on various error metrics.
• Existing literature shows a lack of consensus on the selection and definition of these metrics.
• This ambiguity is especially pronounced in two-dimensional (2D) DoA estimation.

Purpose of the Study:

• To thoroughly investigate and recommend the most suitable error metric for DoA estimation.
• To address the inaccuracies and misleading results from commonly used metrics, particularly in 2D.
• To propose a novel, frame-independent error metric for robust 2D DoA assessment.

Main Methods:

• Comprehensive review of existing DoA error metrics.
• Analysis of metric performance, focusing on 2D DoA scenarios.
• Numerical simulations and experimental validation of the proposed metric against established ones.

Main Results:

• Identified significant inconsistencies and potential inaccuracies in widely accepted DoA error metrics.
• Demonstrated that certain metrics yield misleading results, especially for 2D DoA.
• The proposed error metric provides accurate and meaningful assessments, outperforming others in numerical and experimental tests.
• The novel metric is frame-independent, enhancing its applicability.

Conclusions:

• A specific, frame-independent error metric is recommended for accurate 2D DoA estimation.
• Adopting this metric will improve the reliability and comparability of DoA estimation performance evaluations.
• The findings have broad implications for various practical applications utilizing DoA estimation.