Research on Sound Recognition of UAV Based on Spherical Microphone Arrays

With the popularization of Unmanned Aerial Vehicles(UAV) technology, research on UAV recognition based on sound signals has gradually attracted attention. Traditional microphone array methods usually treat the direction information of the UAV sound source as a prior condition when identifying the sound of UAV. However, in actual environments, the direction of UAV is often unknown, which makes the performance of traditional methods poor in real-world scenarios. This paper utilizes a spherical microphone array combined with the MUSIC (Multiple Signal Classification) and MVDR (Minimum Variance Distortionless Response) algorithms to integrate direction-finding and spatial filtering techniques to achieve precise direction-finding and sound enhancement of the UAV sound source. Compared with a single microphone, the spherical microphone array enhances the sound from the target direction and suppresses environmental noise through the collaborative work of multiple sensors. Compared with traditional array methods that rely on direction priors, this paper combines direction-finding and spatial filtering, effectively adapting to the unknown direction in actual environments. Experiments show that the method proposed in this study has improved recognition rates at different distances (up to 21 meters) for the DJI Air 2S UAV compared to a single microphone, with a 18.98% increase in recognition rate at 21 meters.