Experimental Evaluation of Acoustic Drone Tracking using Mobile Microphone Arrays

With the latest increase in security threats from hostile civil or military UAVs, the need for suitable detection and defense systems is steadily growing. Even small commercial off-the-shelve drones pose a potential security risk when used as a means of attack and can hardly be countered with existing sensor technology. In the development of multi-sensor solutions, the acoustic domain has emerged as a valuable but still underestimated element. Typically, existing acoustic systems focus primarily on the stationary operation and detection or localization of a single drone. However, robust tracking is only achieved if detection and localization components are combined in order to associate a drone sound detection with the respective sound source(s). In this work, we present the development and experimental results of a mobile microphone array system that is able to detect and track multiple drones based on the direction of sound incidence. First, we describe the design and hardware implementation of a semi-spherical array of 32 microphones. Based on our previous work, we then elaborate ways to improve our track-before-detect approach for joint estimation of UAV sound presence and incidence. Finally, we evaluate our proposed system on a dataset acquired in the course of several measurement campaigns for a stationary as well as for a mobile scenario, where the microphone array was mounted on the roof of a vehicle.