Advanced Clutter Mitigation Method for Surveillance Radar using Machine Learning

Abstract

This research focuses on improving the ground clutter mitigation by integrating ML methods with traditional methods (such as CFAR and Doppler processing) of X-band surveillance radar. Discriminative machine learning methods are used as they have the ability to learn without the knowledge of distribution type. The techniques used to accomplish research includes raw IQ radar data collection, data labelling, and feature generation, statistical significance of generated features, model (DT, SVM and ANN) training and model evaluation. The results indicate improvement in mitigation of ground clutter for different scenarios. The research also discusses the future work related to this research.