Dataset and Code for 'Fight Drones with Drones: Detecting Aerial Perimeter Intrusion using Drone-mounted Microphones'

Description This repository contains the dataset and source code associated with the research paper: "Fight Drones with Drones: Detecting Aerial Perimeter Intrusion using Drone-mounted Microphones"Authors: Vincenzo Sammartino, Nathanaël Denis, Omar Ibrahim, and Roberto Di Pietro. Published in the Proceedings of the 31st European Symposium on Research in Computer Security (ESORICS 2026). Overview This project presents an empirically grounded framework for deploying a cooperative drone swarm as an adaptive acoustic barrier to protect critical infrastructure perimeters (e.g., airports, oil refineries, and national borders). Starting from an acoustic sensing and propagation model calibrated on real DJI flight data (Mavic Air 2 and Mini 3 Pro), we analyze the speed–detection trade-off (the Defender's Dilemma) and optimize patrol speeds to minimize the required fleet size while maintaining high detection probability. Contents of this Repository This archive includes: Empirical Calibration Dataset: Telemetry Data: Synchronized 10 Hz flight telemetry (including GPS coordinates, altitudes, 3-axis velocity vectors, and throttle percentage) for the DJI Mavic Air 2 (intruder) and DJI Mini 3 Pro (defenders). Acoustic Recordings: Onboard audio data recorded via RODE Wireless GO II microphones mounted on the patrolling defender drones, covering various flight regimes and pass-by patterns of the intruder. Simulation & Analysis Source Code: Ego-Noise Cancellation Pipeline: Spectral subtraction scripts used to filter out the defender's rotor noise, yielding a significant increase in effective detection range. Monte Carlo Engine: A highly vectorized simulation framework designed to evaluate intruder detection rates across different defender configurations, speeds, and diverse attacker flight profiles (Linear, Zigzag, Evasive, and Altitude-change). Operational Sizing Algorithms: Numerical tools to analyze energy budgets, battery-swap logistics, and collision-avoidance spacing under realistic deployment constraints. Repository & Code Access The active development repository is maintained at:https://github.com/ilsamaritano/BoundaryDefense How to Cite If you find this dataset or code useful in your research, please cite our paper: @inproceedings{sammartino2026drones2D, title={Fight Drones with Drones: Detecting Aerial Perimeter Intrusion using Drone-mounted Microphones}, author={Sammartino, Vincenzo and Denis, Nathana{\"e}l and Ibrahim, Omar and Di Pietro, Roberto}, booktitle={Proceedings of the 31st European Symposium on Research in Computer Security (ESORICS 2026)}, year={2026} }