Search papers, labs, and topics across Lattice.
EchoHawk is an open-source acoustic pipeline designed for the detection, classification, and localization of drones using passive acoustic sensing techniques. The system employs classical wideband beamforming methods and time-delay processing to effectively differentiate drones from low-frequency harmonic confusers, revealing significant performance inflation due to session-level data leakage in existing datasets. By enforcing session-grouped cross-validation, the authors demonstrate a notable reduction in detection performance metrics, emphasizing the importance of proper data handling in acoustic detection studies.
Session-level data leakage can inflate drone detection performance by over 6%, highlighting the critical need for rigorous validation practices in acoustic sensing research.
Passive acoustic sensing is an attractive modality for counter-unmanned aerial system (counter-UAS) defence: it is covert, low-cost, and effective against drones with small radar cross-sections or minimal radio emissions. We present EchoHawk, an open and fully reproducible reference pipeline that detects a drone from its rotor harmonics, estimates its blade-passing frequency, and localises it with a microphone array via classical wideband beamforming (delay-and-sum, MVDR, MUSIC) and time-delay processing (GCC-PHAT, SRP-PHAT), followed by temporal tracking. We evaluate the system on a physically transparent synthetic benchmark that pits drones against hard low-frequency harmonic confusers, such as ground vehicles, and on real recorded audio. Our central methodological contribution is a documented case of session-level data leakage in a widely used public dataset: because its recordings are pre-segmented into short clips, naive clip-level splits place adjacent slices of the same continuous recording in both training and test sets, inflating reported performance. Enforcing recording-session-grouped cross-validation reduces, for example, a random-forest baseline's detection probability at a 1% false-alarm rate from 0.796 to 0.745, yielding honest numbers. All code, figures, and a synthetic data generator are released so that every result runs without any download.
