UAS Detection, Classification, and Tracking in Urban Terrain

Abstract

This paper develops a method for detection, classification, and tracking of small unmanned aerial systems (UAS) in urban environments using a FMCW radar. UAS detection and classification is a major challenge in urban environments due to multipath, low SNRs, and multiple targets such as birds, people, and cars that could trigger a detection. This paper proposes using several features extracted from filtered micro-Doppler signatures to perform binary classification of UAS vs non-UAS targets. Prior to feature extraction and classification a constant false alarm rate (CFAR) detector is used to find potential targets. Post classification, the joint probabilistic data association filter (JPDAF) is used for tracking. Real data results that include several different UAS are presented.