Year: 2025
Type: conference-paper
Abstract: Popular low-cost micro-Unmanned Aerial Vehicles (UAVs) rely on inexpensive MEMS sensors to measure instantaneous rotation and acceleration. These sensors are particularly vulnerable to acoustic injection attacks, which can destabilize the UAV and potentially cause a crash. While prior studies demonstrated the severity of such attacks using expensive equipment in controlled environments (e.g., anechoic chambers), their feasibility for opportunistic, low-budget attackers in realistic operational conditions remains unexplored. In this paper, we experimentally evaluate whether a low-budget attacker can disrupt micro-UAVs' operations via inaudible acoustic injection attacks. Our results show that real-world conditions such as UAV mobility and motor-induced frame vibrations substantially increase attack difficulty compared to the ideal scenario, making it particularly challenging for a low-budget attacker to succeed.
Authors: Jorrit J. Olthuis, Bas Arendsen, Savio Sciancalepore, Nicola Zannone
Institution: Eindhoven University of Technology
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
Award: G5937870420
OpenAlex ID: W7091462322