Target Identification with Improved 2D-VMD for Carrier-Free UWB Radar

Area of Science:

• Radar Systems Engineering
• Signal Processing
• Machine Learning

Background:

• Interest in carrier-free ultra-wideband (UWB) radar for radar automatic target recognition (RATR) is growing.
• UWB radar offers more detailed target information compared to narrow-band systems.
• Real-world environments present challenges due to noise and clutter.

Purpose of the Study:

• To develop and validate a robust RATR system using carrier-free UWB radar.
• To enhance noise reduction and feature extraction techniques for improved target recognition.
• To achieve reliable classification performance in low signal-to-noise ratio (SNR) conditions.

Main Methods:

• Accurate geometric models acquired using 3ds Max.
• Echo signal acquisition via time-domain integral equation (TDIE) for short-duration UWB signals.
• Proposed improved two-dimensional variational mode decomposition (2D-IVMD) for noise elimination and preliminary edge feature extraction.
• Deep conventional neural network (DCNN) for final target recognition.

Main Results:

• Electromagnetic modeling accuracy verified by comparing simulated and actual waveforms.
• 2D-IVMD effectively reduces noise and extracts crucial edge features.
• DCNN achieves promising classification performance.
• The integrated approach demonstrates effectiveness in low SNR environments.

Conclusions:

• The proposed methodology, combining UWB radar, advanced signal processing (2D-IVMD), and deep learning (DCNN), provides a robust solution for RATR.
• The system demonstrates significant potential for accurate target recognition even under challenging noisy conditions.
• This research lays the groundwork for more sophisticated feature extraction and recognition algorithms in UWB radar systems.