Explore the vulnerabilities and potential hacking risks of autonomous vehicles in this informative conference talk from BruCON Security Conference. Delve into the current state of civil driverless vehicles, examining common sensors, decision profiles, and potential failure modes that could be exploited. Gain insights into the US Department of Transportation's Connected Vehicle and Vehicle To Vehicle Communications programs, and their relevance to hacking before the widespread adoption of fully autonomous vehicles. Learn about various autonomous vehicle applications, sensor technologies, and their weaknesses, including GPS spoofing, LiDAR manipulation, and camera-based deep learning vulnerabilities. Discover the implications of these security concerns for the future of transportation and the importance of addressing robustness against adversarial and malicious scenarios in unmanned vehicle development.
Overview
Syllabus
Intro
Retro Talk
Prototype This
Motivation
European driverless vehicle efforts
Autonomous buses
Not here to spread fear
Theyre so good
So this was just some stuff
Another vehicle
Applications
Autonomous Delivery
Civil Applications
Aerial Failures
Sandstorm
Hierarchy
State Machines
Sensors
Sources of uncertainty
Sensor attacks
GPS
Example
GPS spoofing
GPS in civilian use
Lowcost GPS spoofing
GPS spoofing demo
LiDAR
Active Emission Sensor
Transparent Materials
Reflective Materials
Road Line Detection
Solid Object Detection
False Returns
Cameras
Deep Learning
millimeter wave radar
Tesla Autopilot
MEMS Gyros
Failure
Bond vs Robots
The Map
Taught by
BruCON Security Conference
