Breaking the Monopoly of Mechanical Engineering and Dominating with AI: Soongsil University’s 'InTheEND' Team Wins International Autonomous Racing Competition

[Financial News] Students from Soongsil University’s AI College have risen to the top of an international autonomous racing competition with next-generation artificial intelligence technology, breaking through the barrier of traditional mechanical engineering algorithms.
According to Soongsil University on the 1st, the 'InTheEND' team, part of the autonomous driving club AID in the Department of AI Software at the AI College, won the final title at the 29th RoboRacer Autonomous Racing Competition held at BEXCO in Busan.
The victory is especially meaningful because it marks the first time a team has won by putting forward 'AI dedicated to autonomous driving' in a competition long dominated by classical robotics control algorithms based on mechanical engineering.
Kim Kang-hee, a professor at Soongsil University who advised the club, said, "This is the first time a team has won this competition by developing autonomous driving AI, when the mechanical engineering department had always won with robotics algorithms." He added that it was "an achievement that demonstrates the students' outstanding spirit of challenge and the high level of AI technology."
The core technology applied by the students was end-to-end reinforcement learning. Instead of breaking the process into separate steps such as sensing the surroundings with sensors, calculating the driving path, and controlling the steering wheel and speed, they built an AI system that learns driving data and handles everything from situation assessment to control in one step.
Held as part of the 2026 International Federation of Automatic Control (IFAC) event, the competition drew a record 52 teams from universities and research institutes in South Korea, Japan, China, Singapore, India, Thailand, Egypt, and other countries, creating intense technological competition.
The teams built racing cars at one-tenth the size of real vehicles and raced them on the track. In the main round especially, they had to run alongside rival cars on a course with obstacles, avoid unexpected situations in real time, and maintain speed, demanding top-level control technology.
Soongsil University entered nine teams in the competition, and six advanced to the round of 32. Among them, the InTheEND team passed the preliminaries in second place and maintained overwhelming performance through the finals to lift the trophy.
The winning team members said, "We faced many difficulties at the beginning as we prepared for the competition using end-to-end reinforcement learning, which was unfamiliar to us, starting in December last year. But by working together with our teammates, we achieved a rewarding result." They also expressed gratitude to their advisers and to the AI Advanced Industry Talent Development Boot Camp project team, which provided generous support.  

[email protected] Kim Man-ki Reporter