Classification of Fighter Jets with Radar HRRP Data and Machine/Deep Learning Methods

The classification of air targets is crucial in air traffic control, border security, platform defense, and tactical picture compilation. In addition to using radar-derived target dynamic information among other sensors, high-resolution range profiles (HRRP) can provide significantly more comprehensive information about the target’s shape and dimensions for classification. In addition to being able to efficiently and swiftly interpret the information acquired by radar systems, it is also necessary to generate the data required for classification via measurement or digital simulation. This study compares fighters classification performance using machine learning and deep learning on HRRP data. For this goal, many backbones based on the basic infrastructure of convolutional neural networks (CNN) were investigated, including Alexnet and Resnet50. Following feature extraction utilizing deep learning approaches, classification results were shared using support vector machines (SVM) for machine learning. Furthermore, classification investigations were carried out using various feature extraction approaches and machine learning techniques. The classification results were evaluated using metrics such as precision, recall, and the F1-score. The classification results for the selected sample fighters were calculated and shared.