Dronar: Obstacle Echolocation Using Drone Ego-Noise
Henrik Nilsson, Joakim Rydell, Anton Kullberg, Gustaf Hendeby · 2024
A method for obstacle detection using the sound that a drone naturally emits is proposed. The sound emitted from a vehicle, ego-noise, is often considered a complicating factor for mission fulfilment, without purpose. The idea in this paper is to utilise this ego-noise for obstacle detection, being the first to perform practical experiments of this. Adding a few microphones to the vehicle, the ego-noise is utilised as the sound source for echolocation. The method consists of auto-correlating the received signals to estimate echo delays, using the known array geometry and signal propagation speed to relate delays to distances, and then beamforming to position targets. A proof-of-concept has been constructed, and promising results are presented for experiments in a controlled environment.