Defending Against Consumer Drone Privacy Attacks: A Blueprint for a Counter Autonomous Drone Tool

Abstract

Drone technology has become ubiquitous, touching almost every aspect of our society. Thus, it is no wonder that malicious actors have now expanded their selfish and criminal pursuits by using drone technology. Currently, these actors commit terrorist acts internationally and domestically, they commit corporate espionage or transport drugs or contraband, and they even violate the privacy of their own neighbors. These are just a few of the many misguided uses of rogue drones. Many researchers in government, corporate, and academia have proposed viable solutions to mitigate the risk associated with this current threat. However, most of these solutions focus on tracking and disrupting the Wi-Fi link between the user and the drone, which may be sufficient for the current threat from Wi-Fi drones, but the future threat is rogue autonomous drones. The use of autonomous drones for rogue purposes poses a more difficult threat to counter since these drones navigate on their own (once given a mission), may not use Wi-Fi at all, and do not require a human in the loop. Given these autonomous drone characteristics, existing user-controlled drone countering tools are ineffective against them and to our knowledge, no other low cost, non-military, and effective counter autonomous drone tools exist. Using DJI drones, which are 74% of the global commercial drone market, in autonomous mode (Active Track), we demonstrate how the use of "hard to patch" vulnerabilities, which break the drone's autonomy code, can be weaponized to design an effective counter autonomous drone tool capable of mitigating one or multiple drones.

Countering a Drone in a 3D Space: Analyzing Deep Reinforcement Learning Methods

“…Researchers Watkins et al proposed a blueprint [14] which offers a design for a novel autonomous counter drone tool based on the weaponization of "hard-to-patch" vulnerabilities. The paper highlights the problem of privacy violation due to drones and presents a counter-drone tool which breaks the drone's autonomy code.…”

Counter a drone via deep reinforcement learning

“…It has been demonstrated that the platform is capable of detecting drones, determining their coordinates, and neutralizing them using a "hunter" drone in order to accomplish the intended tasks. Also, Watkins et al (2020) illustrated how to design a powerful counter autonomous drone tool that can be used against a single drone or a group of drones by using hard-to-patches vulnerabilities. Authors claim that this counter autonomous drone tool fills a critical need for mitigation of risks due to privacy violations such as recording video by drones.…”