Detection of Drones using Acoustic Signals and Machine Learning

Drone detection using radar is popular. Diffraction effect of the signal is the main challenge in radar-based detection. An alternative approach to address the diffraction effect is acoustic signals. In this work, acoustic based detection of drones is proposed. Various possibilities of drones are single rotor, multi rotor, fixed wing, nano, medium, large and heavy lift. Acoustic signals are different for various types of drones. Features from acoustic signals are integrated with machine learning algorithms. Pitch features are important in audio signals. Hence, in this paper pitch and other features are explored to detect drones. To classify the drones, Decision Tree and Support Vector Machine based Machine Learning algorithms are explored. It is easy to detect drones using acoustic signals instead of radio frequencies, cameras, and radars. Performance of the proposed methods are assessed using confusion matrix.