Navigating Chaos: The Consequences of Compromise in Mixed-Autonomy Freeways
summary
Compromised Adaptive Cruise Control modules accept injected CAN commands and, with a simple randomized deceleration attack, produce stealthy micro‑brakes that amplify into wide congestion in high‑fidelity simulations and a controlled road test. A lab record‑and‑replay of 30,000 CAN packets and a supply‑chain ECU implant both succeed at forcing those inputs, while AutoEncoding Neural Networks trained on benign GPS trajectories flag anomalous vehicles but struggle with noisy, mixed‑autonomy traffic.
AI-written summary · about
chapters (7)
- 0:00Mixed-autonomy traffic threats
- 6:35Vehicle automation levels
- 15:06CAN bus traffic disruption
- 17:58Randomized deceleration attack
- 22:08Detecting compromised vehicles
- 36:37Mitigations and implications
- 39:52Q&A
related talks
50mAdvanced CAN Injection Techniques for Vehicle Networks
13mWIP: Learning Adversarial Attacks on Adaptive Traffic Signal Control Systems...
14mStateful Behavior Inference and Runtime Enforcement for Vehicle Network Security
43mConnected Cars - What could possibly go wrong?
09mLightning Talk: On-Road Driver Identification Dataset with CyberAttack
13mAUTO: Adaptive Congestion Control Based on Multi-Objective Reinforcement Learning...
37mAutomatic Automotive Hacking
06mWhen Collaboration Betrays: The Evolving Attack Surface of...