Original title: Mitigating Degradation Attacks in Cooperative Autonomous Driving via Intention Sharing: A Vehicle-in-the-Loop Study
Prakhar Gupta, Tyler Ard, Rongyao Wang, Jagruti Sahoo +3 · September 2026
Connected self-driving cars rely on wireless messages to travel safely together. However, cyber attacks can cause major communication delays that lead to crashes. Researchers wanted to see if sharing planned future moves could solve this danger. They tested cars using a setup that combines a real drive-by-wire car, simulated traffic, and wireless vehicle links. They simulated attacks that delayed messages by up to two seconds. Instead of only broadcasting their current status, cars shared their future intentions with each other. This extra foresight lets vehicles plan around delayed data. The researchers found that cars sharing only their current status suffered severe performance drops and crashed often. In contrast, intention sharing eliminated collisions entirely and kept driving behavior near normal levels across the tested scenarios.
You are driving in a convoy and need to coordinate turns. Normally each driver signals their current speed and position, but if walkie-talkies are jammed by interference, messages arrive late and you might crash. Here, instead of reporting current status, each driver announces their intended next few moves, so even if messages are delayed you can adjust your plan to avoid collisions.
Connected autonomous cars are vulnerable to cyberattacks that delay communication. This research shows that by sharing their planned actions instead of just current status, these cars can avoid collisions even when attacked. This could make self-driving fleets safer against real-world cyber threats.
The study is limited to specific scenarios and communication delays up to two seconds. It does not address other attack types, longer delays, or broader driving conditions. The vehicle-in-the-loop testbed uses one real vehicle, so results may differ in fully deployed systems.
Ask anything about this paper. Answers come from its abstract, and say so when the abstract doesn't cover your question.
Mitigating Degradation Attacks in Cooperative Autonomous Driving via Intention Sharing: A Vehicle-in-the-Loop Study — Prakhar Gupta, Tyler Ard, Rongyao Wang et al.
Abstract licence: CC0
Explanation generated by Synthesis from the abstract — it is not the original paper. This is a preprint, so it hasn't necessarily been peer reviewed.