Radar Target Detection Based on Linear Fusion of Two Features

Area of Science:

• Radar systems engineering
• Signal processing
• Target detection

Background:

• Multi-feature detection improves radar capabilities for weak sea surface targets.
• High-dimensional data and lack of robustness limit current multi-feature detection methods.

Purpose of the Study:

• To propose a radar target detection method reducing feature dimensions via linear fusion.
• To enhance detection performance and applicability for weak sea surface targets.

Main Methods:

• Developed a two-feature linear dimensionality reduction method based on distribution compactness.
• Fused features were modeled using the generalized extreme value (GEV) distribution.
• Designed an asymptotic constant false alarm rate (CFAR) detector based on the GEV distribution.

Main Results:

• The proposed method effectively reduces feature dimensions.
• The GEV distribution accurately models the fused feature's tail probability density function.
• The developed CFAR detector demonstrates robust detection performance.

Conclusions:

• Linear fusion of two features offers a viable approach for radar target detection.
• The GEV-based CFAR detector improves detection of weak sea surface targets.
• The method addresses limitations of high-dimensional data and robustness in radar systems.