WIP: Hijacking Attacks on UAV Follow-Me Systems in Realistic Scenarios
summary
A gray‑box attack pipeline runs real‑time 3D detection + gradient optimization on a high‑fidelity simulator geo with an open‑source autopilot (PX4) to project 3D positions into the gimbal camera view and steer a surrogate multi‑object tracker; iterating this loop raises hijack success from 47% to 95% success by exploiting sensor noise and gimbal instability, and acoustic injection induces >20° oscillation on a DJI gimbal to materially increase real‑world attack effectiveness.
AI-written summary · about
related talks
13mBlind Spot in Sensor Fusion: Exploiting an Architectural Vulnerability to Hijack...
20mSLEIGHT: Real-time Closed-Loop Attack Analytics Framework for Holistic Intrusion...
28mHacking a Professional Drone
21mInjected and Delivered: Fabricating Implicit Control over Actuation Systems by Spoofing Inertial Sensors
40mSecuring Tesla and Mobileye from Split-Second Phantom Attacks
01h 06mSelf-Driving and Connected Cars: Fooling Sensors and Tracking Drivers
19mAirways: Optimization-Based Planning of Quadrotor Trajectories according to High-Level User Goals
12mHacking Drone Warfare