Korea Univ.-Kookmin Univ. joint research team wins top vehicle security award

Members of the joint research team from Korea University’s Hacking and Countermeasure Research Lab and Kookmin University’s Mobility Cybersecurity Lab receive the Best Artifact Award at USENIX VehicleSec 2026, a leading international academic forum for vehicle cybersecurity, held in Baltimore, Md., Tuesday. Courtesy of Korea University

A research team jointly led by Korea University’s Hacking and Countermeasure Research Lab (HCRL) and Kookmin University’s Mobility Cybersecurity Lab has received the Best Artifact Award at USENIX VehicleSec 2026, the universities said Wednesday.

The two-day international symposium, regarded as a leading academic forum for vehicle cybersecurity, was held in Baltimore, Maryland, earlier this week.

The recognition comes as automakers increasingly adopt advanced and connected technologies, raising concerns over vehicle hacking and increasing demand for artificial intelligence (AI)-powered systems capable of detecting cyberattacks targeting in-vehicle networks.

The team was recognized for developing the AutoHack Dataset, a vehicle cybersecurity dataset collected from cars during actual road driving. The dataset is designed to help AI-based security systems distinguish between normal vehicle activity and malicious attacks.

The findings were detailed in a paper titled “AutoHack: A Physically Verified Multi-Bus CAN Dataset for Intrusion Detection System Evaluation.”

The paper received all three badges for openness, functionality and reproducibility in the conference’s artifact evaluation and went on to win the Best Artifact Award, recognizing the quality and reliability of the research.

Most existing vehicle cybersecurity datasets are collected in simulated environments or from a single in-vehicle network, making them less representative of real-world driving conditions, where unexpected and irregular communications frequently occur.

By contrast, the research team collected data from three in-vehicle networks simultaneously and synchronized the traffic at the microsecond level. This allowed the researchers to examine how the networks interact and how an attack can spread between them.

The dataset was also tested in AutoHack 2025, Korea’s only vehicle hacking and defense competition using real cars, demonstrating its practical applicability.

The dataset and benchmark code are freely available through Zenodo, an open platform for sharing research materials.

“This achievement will help create a cycle in which cybersecurity competitions lead to the release of useful datasets and encourage more researchers to conduct follow-up studies,” said professor Kim Huy-kang, who leads Korea University’s HCRL.