A Novel Neural Network Architecture for Radar Clutter Classification

In recent years, potential capabilities of modern radars have become tremendous, with rapid advances in electronics and software technologies. If the radar system processes received echoes based on clutter characteristics, detection performance significantly improves. To classify clutters, neural network structures are studied. A problem specific architecture, specialized in clutter classification, is designed. The design procedure of the neural network is explained with necessary theoretical background information. The performance of the network is illustrated with experimental results.