LFM pulse radar-based uav detection and classification using machine learning and data augmentation to improve performance

Abstract

In recent years, the rapid technological development and ease of use of unmanned vehicles have led to their application in various fields and a growing market. However, due to the damage caused to private spaces and attacks on nationally important facilities by unmanned vehicles, the importance of detecting and classifying unmanned vehicles has increased. Over the past decade, there have been many studies on detecting and classifying unmanned vehicles. In this thesis, a mathematical model of the characteristics of unmanned vehicles is designed and a classification technique using it is presented. First, we present an algorithm for classifying unmanned vehicles and birds using support vector machines. It also presents an algorithm for the classification of unmanned vehicles and their distinction from birds using CNNs. Finally, we discuss the challenges of substation environments for UAV detection and suggest research needs.