Kookmin University research team’s paper accepted for presentation at international robotics conference

Undergraduate student Kim Jun-ho, left, and professor Lee Seong-won from Kookmin University’s School of Electrical Engineering / Courtesy of Koomin University

A Kookmin University research team has developed a new technology that helps artificial intelligence (AI) better understand 3D environments when objects are hidden from view or visual information is limited.

The university said Friday that a paper detailing the technology has been accepted for presentation at the 2026 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) set for Sept. 27 to Oct. 1 in Pittsburgh.

The team was led by Lee Seong-won, professor at the university’s School of Electrical Engineering. The team included Kim Jun-ho, a senior majoring in electrical engineering.

Kim is the first author of the paper titled “RayOcc: Occlusion-Agnostic Ray Occupancy Estimation via Gaussian Mixture Intensity,” which proposes a robust AI-based method for estimating the occupancy of 3D spaces, helping overcome visibility limitations in complex environments.

The university noted that the technology addresses challenges in 3D spatial understanding — essential for autonomous driving and robotics systems — and demonstrates the potential to overcome uncertainties caused by limited observations and hidden areas in real-world environments.

It expects the technology to make broad contributions to the fields of autonomous driving and intelligent robotics in the future.

The emblem of the 2026 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) to be held in Pittsburgh from Sept. 27 to Oct. 1 / Courtesy of Kookmin University

The research was conducted at the university’s Computer Intelligence Lab under Lee's guidance.

Undergraduate student Kim played a leading role in all aspects of the project — from research design to implementation and experimentation.

The university said the study is significant as it represents a case where an undergraduate student participated throughout the entire research process and had the paper accepted for presentation at an international academic conference.

Kim also served as a lead author on a paper titled “VG3T: Visual Geometry Grounded Gaussian Transformer,” which was accepted at the 2026 IEEE International Conference on Robotics and Automation (ICRA) held in Vienna from June 1-5.

The university said Kim’s consecutive acceptance of lead-author papers at major international robotics conferences demonstrates the university’s growing research capabilities.

“I find it very meaningful that the research I took part in as an undergraduate has been recognized at IROS following its acceptance at ICRA,” Kim said.

"Building on my experience in research design, implementation and experimentation, I would like to continue conducting AI research that can contribute to real-world robotic systems and autonomous driving applications.”