Audio-Based Measurement of PWM-Encoded Spectral Features in UAV Acoustic Emissions

The widespread use of aerial vehicles, both manned and unmanned, can raise issues, especially around critical civil infrastructure. Unauthorized drones should be immediately identified and forced to land or be defeated. Although traditional radio frequency (RF)-based methods are well-established, other unconventional methods of identification are also interesting to explore and study, as they can be more robust and less prone to jamming in hostile situations. In this work, a communication method enabled by audio measurements is proposed. A customized Pulse Width Modulation (PWM) signal is used to control the drone's motors, and in turn its propellers, modulated by a designed signal. Measurements of the propeller acoustic emissions, combined with a specific signal processing technique, demonstrate the feasibility of the approach, meaning that such a modulation of the driving signal, causing variations in the acoustic noise emitted by the drone, can be leveraged to encode, and successively decode, communication symbols with which establishing intrinsically secure communications over a so-called covert channel.