Projection Metric Learning of Updated-Subspaces for Radar Target Classification

In this study, we focus on the classification problem of small radar targets. Persistent monitoring of a radar target requires an algorithm with low-computational cost and high-accuracy classification. To achieve this, we propose a radar based UAV classification framework that integrates signal detection by a subspace-updating method with projection metric learning on a Grassmann manifold. The proposed method was demonstrated through experiments using a five UAVs data set collected from continuous wave (CW) radar and exhibiting up to 4.28 points higher-accuracy and 93 times faster than the conventional method.