Classification Algorithm for Radar

To address the Army’s need for advanced target classification algorithms for ground-based multi-mission radar (MMR), Physical Optics Corporation (POC) proposes to develop a Classification Algorithm for Radar (CAR), which provides a high classification capability using target features to discriminate among unmanned aerial systems (UASs), manned aircraft, and clutter with very high confidence, as well as assist in precise trajectory prediction. CAR includes proprietary parametric models, and classifiers that allow CAR to distinguish unmanned aerial vehicles (UAVs), non-UAV objects, fixed-wing and rotary-wing UASs, and clutter; further separate them into target type and sub-type categories; and enhance the trajectory prediction of mobile targets with very low false positive rates. CAR, when integrated with future Army air surveillance and MMR systems such as the AN/TPQ-50 and AN/TPQ-53, will not just provide an advanced classifier that enhances the performance of current radar, but also ensure radar resources are not wasted in identifying false target tracks; hence, prediction time for identifying a target is much faster in CAR. In Phase I, POC will demonstrate the feasibility of CAR by investigating various feature extraction algorithms and various classifier functionalities. In Phase II, POC will develop a field-ready CAR classifier with supporting algorithms.