KIMM Fast Tracks AI System, Robotics Development

Korea Institute of Machinery and Materials (KIMM) has developed a robot task AI system designed to learn and perform everyday repetitive tasks through human demonstrations.
The technology, led by Dr. Jeong-Jung Kim, facilitates robots in completing common activities like organizing, clearing, and manipulating objects in various environments.
At its core, the system uses a hierarchical task execution framework to break down complex jobs into sequential steps for systematic completion.
The platform integrates three key technologies: task extraction, virtualized training environments, and hierarchical execution AI.
Unlike many existing systems that rely on single-task datasets or simulation-only testing, KIMM’s framework includes the full development pipeline—from dataset construction to real-world physical robot testing.
During testing, the system achieved above 90% success rates in multiple tasks, showing strong reliability and adaptability under changing conditions.
The new technology is expected to boost the role of service robots in sectors like retail merchandising, warehouse logistics, and general workplace support.
The research is part of the KIMM RoGeTA (Robot General Task AI) framework program, part of which involves releasing datasets and virtual models to support future service robot developments.

You may have missed