A Machine Learning Approach to Helicopter Open set Classification based on Microdoppler Signatures

Detecting and identifying rotorcrafts is a critical challenge in the domains of defense and security. Helicopters, due to their low speed and altitude, pose unique difficulties in radar-based detection. Moreover, accurate identification is strategically important. Thankfully, helicopters’ radar return signals are modulated with characteristic frequencies directly related to the rotary motion of their rotors. Leveraging this micro-Doppler radar signature, in addition to existing sensors, can significantly enhance the detection and classification performance. In this article, we explore the application of neural network-based algorithms to identify rotorcraft in an open-set context, accounting for scenarios where samples could belong to unknown classes as in an operational scenario.