Synthetic Imaging Radar Data Generation in Various Clutter Environments Using Novel UWB Log-Periodic Antenna

Area of Science:

• Electromagnetics
• Microwave Imaging
• Artificial Intelligence/Machine Learning

Background:

• Collecting real-world data for short-range microwave imaging target detection is labor-intensive and requires large datasets for AI/ML development.
• Existing methods often struggle with cluttered environments, necessitating alternative data generation approaches.

Purpose of the Study:

• To explore synthetic data generation using electromagnetic wave propagation simulations as a cost-efficient complement to experimental data.
• To develop and validate a novel ultra-wide band (UWB) log-periodic antenna for realistic synthetic data generation.
• To assess the utility of generated synthetic data for target detection applications like through-the-wall and through-the-foliage imaging.

Main Methods:

• Designed and simulated a 3-D model of a novel printed scalable UWB log-periodic antenna with a tapered feed line.
• Incorporated the antenna model into electromagnetic wave propagation simulations to generate synthetic microwave imaging data.
• Generated synthetic data for Through-the-Wall Imaging (TWI) and Through-the-Foliage Imaging (TFI) scenarios.
• Explored and analyzed clutter removal techniques on the generated synthetic data.

Main Results:

• The designed UWB log-periodic antenna exhibits a highly directional radiation pattern and high gain (>6 dBi) across its bandwidth.
• Synthetic data generated using the 3D antenna model realistically captures environmental coupling effects.
• The synthetic imagery proved suitable for developing and training machine learning models for TWI and TFI applications.

Conclusions:

• Synthetic data generation via electromagnetic simulations offers a viable and efficient alternative to experimental data collection for microwave imaging.
• The proposed UWB log-periodic antenna is effective for creating realistic synthetic datasets for target detection.
• The synthetic data facilitates the development of AI/ML models for challenging imaging applications like TWI and TFI.