Automatic target recognition based on receiver optimization of photonic time-stretched coherent radar

Area of Science:

• Radar Systems Engineering
• Signal Processing
• Artificial Intelligence

Background:

• Automatic target recognition (ATR) is crucial for defense and surveillance.
• Photonic time-stretched coherent radar (PTS-CR) offers advantages in signal processing.
• Existing PTS-CR systems face limitations in reception aperture and delay information restoration.

Purpose of the Study:

• To develop an improved PTS-CR scheme for enhanced ATR capabilities.
• To increase the reception aperture and accurately restore signal delay information.
• To validate the scheme's performance through simulation and experimentation.

Main Methods:

• Implemented an improved PTS-CR receiver with high repetition rate pulses.
• Utilized two channels with different stretching factors to restore signal delay.
• Integrated the enhanced PTS-CR with a convolutional neural network (CNN) for target classification.

Main Results:

• The improved PTS-CR reception aperture covers the entire detection range.
• Simulated and experimental results confirmed the feasibility of the new scheme.
• Achieved a 94.375% ATR accuracy rate on four distinct targets using the CNN.

Conclusions:

• The enhanced PTS-CR scheme effectively improves reception capabilities for ATR.
• The integration of PTS-CR with CNNs provides a robust solution for target recognition.
• This approach demonstrates significant potential for advanced radar-based surveillance and identification.