PCA-Based Matrix CFAR Detection for Radar Target.

Area of Science:

• Radar signal processing
• Information geometry
• Statistical detection theory

Background:

• Constant False Alarm Rate (CFAR) is crucial for radar target detection.
• Matrix CFAR detection utilizes Hermitian positive-definite (HPD) covariance matrices for efficient detection.
• HPD matrices suffer from information redundancy, limiting detection performance.

Purpose of the Study:

• To propose a Principal Component Analysis (PCA) based matrix CFAR detection method.
• To address point target detection challenges in clutter environments.
• To enhance detection performance by reducing dimensionality and redundant information.

Main Methods:

• Applying PCA to the cell under test to reduce HPD covariance matrix dimensionality.
• Constructing a transformation matrix for mapping to a lower-dimensional space.
• Deriving detection statistics and decisions on the matrix manifold.

Main Results:

• Reduced dimensionality and redundant information in HPD covariance matrices.
• Enhanced distinguishability between targets and clutter.
• Improved signal-to-clutter ratio (SCR).

Conclusions:

• The proposed PCA-based matrix CFAR method significantly improves radar target detection performance.
• The method is validated through simulation and real sea clutter data experiments.
• PCA effectively mitigates information redundancy in HPD covariance matrices.