A UNIST autonomous racing team has won first place at RoboRacer, an international autonomous driving competition, held as part of the 2026 IEEE International Conference on Robotics and Automation (ICRA), one of the world's leading conferences in robotics and automation.

Team UNICORN Racing , led by Professor Cheolhyeon Kwon of the Department of Mechanical Engineering, claimed the championship title at the competition, which took place in Vienna, Austria, from June 1 to 5, 2026. The team finished ahead of 37 university teams from around the world, demonstrating the strength of its autonomous racing systems under demanding real-world conditions. 

RoboRacer, formerly known as F1TENTH, challenges teams to develop perception, planning, and control systems for autonomous one-tenth-scale race cars operating under high-speed racing conditions. 

This year's course introduced significant new challenges. Unlike a conventional flat track, the racing environment included ramps, bridges, and elevation changes that caused vehicles to lift off, land, and experience sudden changes in posture. Success depends not only on speed, but also on the vehicle's ability to perceive its surroundings, accurately estimate its position, and maintain stable control under rapidly changing conditions.

Team UNICORN Racing distinguished itself with a technical approach not used by any other team in the competition that was a three-dimensional (3D) LiDAR system. While most teams relied on conventional two-dimensional LiDAR sensors, the UNIST team used 3D LiDAR to capture richer spatial information about the racing environment.

The approach addresses a broader challenge in autonomous systems research. In complex three-dimensional environments, changes in terrain and vehicle motion can disrupt perception and localization, making reliable navigation difficult. Processing large volumes of three-dimensional sensor data in real time also requires substantial computing power, particularly on small autonomous platforms with limited onboard resources.

To overcome these challenges, the team developed an integrated system that combines 3D perception, localization, path planning, and vehicle control within a compact computational framework. The system maintained accurate localization and stable control even on sections with sharp elevation changes and landing impacts, enabling consistent performance throughout the race and helping secure the overall victory.

“This achievement is especially meaningful because it builds on last year's second-place finish and demonstrates the continued progress of our autonomous driving research,” said Professor Kwon. "Our goal is to advance technologies that enable fast and reliable autonomous driving in complex three-dimensional environments."

RoboRacer began at the University of Pennsylvania in 2016 and has since grown into a global research and education platform involving more than 100 universities worldwide. By combining real-world racing conditions with autonomous systems research, the competition provides a demanding testbed for technologies that may inform the future of intelligent mobility.