Kakao Mobility has released an AI learning dataset for autonomous driving research specifically tailored to South Korean road conditions, removing a key barrier for smaller firms and researchers developing self-driving technology.
The dataset, announced Thursday as part of a national initiative led by South Korea’s Ministry of Science and ICT, includes 150,000 entries collected from major domestic roadways. It captures both 3D dynamic objects like pedestrians and vehicles, and 2D static elements such as traffic signs, using LiDAR and camera sensors mounted on Kakao’s autonomous vehicles.
Prior to this release, many researchers in Korea were forced to rely on foreign datasets that didn’t accurately represent local driving conditions, according to the company. The new resource covers 31 categories across various road types, weather conditions, and times of day.
“This dataset will serve as a cornerstone for accelerating the commercialization of domestic autonomous driving technology,” said Jang Sung-wook, head of Kakao Mobility’s Future Mobility Research Center.
The collection is now available through the Electronics and Telecommunications Research Institute’s AI Sharing platform. It supports the national project’s goal of advancing Level 4 autonomous driving, where vehicles can operate without human intervention under specific conditions.
Kakao Mobility’s initiative joins a growing trend of open-data sharing in the autonomous vehicle sector, as companies and governments worldwide recognize that broader access to training data can accelerate development timelines.
