Distinguishing Drone Types Based on Acoustic Wave by IoT Device

摘要

A drone detection system is essential to protect people from misuses, infringement on privacy, and other harms by drones. In this paper, we investigate to distinguish drone types based on acoustic wave. Our proposed drone detection technique employs Dejavu that is an acoustic fingerprint identification tool and random forest to classify types of drones. In the study, we first collected a dataset for three types of drones: DJI Spark, Quadcopter AR Drone 2.0, and Parrot Mambo drone. Then, we compared the detection performance based on the dataset. The experimental results showed that each drone sound can be classified by random forest.