Jeonbuk, South Gyeongsang Launch KRW 1.4131 Trillion Physical AI Project

A physical AI research and development project centered on Jeonbuk Special Self-Governing Province and South Gyeongsang Province has officially gotten under way. The total project cost is KRW 1.4131 trillion. The project is structured so that Jeonbuk and South Gyeongsang each handle different areas.

Jeonbuk will be responsible for building a physical AI software (SW) platform linked to manufacturing and logistics sites and for creating a demonstration complex. South Gyeongsang will handle the development of a large action model (LAM) specialized in precision manufacturing processes.

The Ministry of Science and ICT and the National IT Industry Promotion Agency (NIPA) held a kickoff briefing for the Jeonbuk-South Gyeongsang physical AI research and development project on the 16th at the Novotel Ambassador Seoul Dongdaemun. At the briefing, detailed implementation plans through 2030 were shared.

The project focuses on linking manufacturing data, AI models, digital twins, robots, equipment control, and field verification into a single research and development framework. It also includes a plan to apply technologies developed in the lab to manufacturing sites and to reuse data secured on-site to train AI models and upgrade operating systems. Through this, a cycle connecting research and development, field application, and data feedback was proposed.

Jeonbuk will pursue this year through 2030 the project titled 'Research and development of a collaborative-intelligence physical AI-based SW platform and ecosystem development.' A total of KRW 736.8 billion will be invested in the project, and KRW 515 billion including government funding. Jeonbuk will carry out the project by developing core physical AI technologies specialized for manufacturing and logistics, along with a manufacturing and logistics-oriented SW platform, and by promoting verification in actual industrial settings.

To that end, it will form a consortium centered on Jeonbuk National University and CAMTIC Advanced Institute of Technology to develop a platform that enables information sharing and collaboration among different robots and equipment. It also plans to create a demonstration complex in the Iseo-myeon area of Wanju County. The complex will include a joint research and development center, an NPU technology verification station, and a demonstration meta-factory facility, and the project will push ahead with building a foundation linking research and development, on-site verification, and commercialization.

South Gyeongsang will pursue a separate project, 'Human-AI collaborative LAM development and global demonstration.' The total project cost is KRW 676.3 billion, with universities, research institutes, and manufacturing companies participating. Seoul National University, Gyeongsang National University, LG AI Research, and Hanwha Ocean will join as participating institutions, and they will pursue the development of physical AI technologies specialized for precision manufacturing processes and global demonstrations.

LAM is presented as a core technology that makes actual action execution and precision control possible based on data combining manufacturing processes, robot behavior, physical laws, and spatial information. To this end, a build framework that links digital twins, data pipelines, simulations, and neural processing unit (NPU)-based AI infrastructure is needed, along with a circular system ranging from data training to field deployment and retraining based on results.

South Gyeongsang plans to build an integrated platform and key verification systems this year and next, including representative manufacturing process data, a core data pipeline, and a bidirectional digital twin. It then aims to develop a LAM through 2030 that understands the physical laws of manufacturing processes and performs ultra-precise control, and has set verification of autonomous control for manufacturing equipment and robots as a target.

The government plans to secure representative manufacturing process data, a data pipeline, a bidirectional digital twin, a platform minimum viable product (MVP), core technologies, and verification systems by next year. It also aims by 2030 to build digital twins and virtual factories for more than 13 manufacturing processes and, in integrated demonstrations, to achieve a closed-loop automation rate of 90% or more and retraining reflection within 24 hours.

Jeonbuk and South Gyeongsang share the common goal of using manufacturing sites as a stage for verifying physical AI technologies, but their project directions differ. Jeonbuk is focusing on a SW platform and demonstration infrastructure that connect various robots, manufacturing, and logistics equipment, while South Gyeongsang is focusing on developing a LAM that directly controls equipment and robots in precision manufacturing processes with AI.

Lee Do-gyu, director of the Ministry of Science and ICT's Office of ICT Policy, said the ministry will support the spread of Jeonbuk and South Gyeongsang's technology development and on-site demonstration achievements to manufacturing sites nationwide. He also said this will ultimately support new industrial competitiveness that could lead to exports of South Korea's 'K-manufacturing factories' to global markets.

Source: TECHWORLD · Kim Seung-gi
Original: https://www.epnc.co.kr/news/articleView.html?idxno=407077

References

• https://www.epnc.co.kr/news/articleView.html?idxno=407077