Konan Technology Partners With Youngpoong Electronics to Develop Manufacturing and Defense AI Agent
TECHWORLD ·
✦ AI Summary
Konan Technology has been selected for the Ministry of SMEs and Startups and TIPA-run "Fast Commercialization Support Project for AI Application Products in Manufacturing" and has begun work on the "AI-based manufacturing knowledge assetization and automated management of electronic circuit board manufacturing processes" project.
The project will run through July 2027, and total project costs are KRW 1,800,008,000.
The participating organizations will build the failure diagnosis know-how of experienced workers in defense PBA manufacturing into an ontology-based knowledge graph so that AI agents can assist with failure investigation, document writing, and technical document Q&A.
Konan Technology said on the 31st that it had been selected for the Ministry of SMEs and Startups and the Technology and Information Promotion Agency for SMEs (TIPA)-run "Fast Commercialization Support Project for AI Application Products in Manufacturing" and had begun work on the "AI-based manufacturing knowledge assetization and automated management of electronic circuit board manufacturing processes" project. The project will run through July 2027, and total project costs are KRW 1,800,008,000.
Under this project, Konan Technology and Youngpoong Electronics will cooperate on developing manufacturing and defense AI agents. Youngpoong Electronics, a defense electronics components company, will participate as the adopting company, while Konan Technology will be responsible for supplying the AI solution.
Abercus will take part in developing the AI agent, and the Industry-Academic Cooperation Foundation of Changwon National University will be in charge of ontology and knowledge graph design. Gyeongnam Technopark will oversee overall project planning.
The participating organizations will advance the development of ontology-based AI agents with the goal of codifying the know-how of experienced workers on defense manufacturing sites and supporting failure diagnosis and process tasks.
The core of this project is to build, as an ontology-based knowledge graph, the failure diagnosis experience and know-how accumulated by experienced workers in the manufacturing process of defense PBA assembly. The idea is to structure experts' knowledge so that AI can assist with failure investigation tasks.
Current failure investigation work is carried out in a high-mix, low-volume production environment and relies on individual experience and skill levels. Accordingly, the AI support aims to help junior engineers with diagnosis tasks and support the transfer of skilled expertise.
The development targets include an inspection agent that suggests failure causes and suspect components by analyzing schematics and datasheets. It also includes an agent that searches for similar cases based on past failure histories.
Among the functions to be built is the creation of failure investigation reports using LLM. In addition, a RAG-based technical document Q&A function will also be provided.
The operating model is Human-in-the-Loop, in which AI performs a first-pass analysis of failure causes and engineers verify and approve the results. The functions will cover circuit and component information analysis, past case searches, report writing, and technical document Q&A, but final judgment will rest with engineers.
Through this project, Konan Technology aims to reduce failure-cause analysis time by more than 50% and report-writing time by more than 70%.
Konan Technology sees this project as an extension of its own defense "AI staff officer" and defense and manufacturing "AI master" strategies. It is envisioning the expansion of AI technologies secured in the defense sector to manufacturing sites, with the expansion targets being the assetization of skilled knowledge and process tasks.
Kim Tae-hyung, head of Konan Technology's Defense and Strategic Business Division, said the representative pain point on manufacturing sites lies in the persistence of experienced workers' tacit individual know-how, and added that the company will begin full-scale efforts to assetize that knowledge and turn manufacturing and defense AI agent applications into "AI master" solutions. He also said the goal is to solve productivity and skills-transfer issues at the same time.
Konan Technology is also pursuing a plan to commercialize an on-premises B2B solution for internal enterprise operations based on a combination of LLM, RAG, and graph RAG. After the proof of concept is completed, it plans to expand the target users to other defense electronics manufacturers in the same industry.
Source: TECHWORLD · Kim Seung-gi
Original: https://www.epnc.co.kr/news/articleView.html?idxno=406261
References
This article was produced with the help of an automated content generation algorithm.
Source: TECHWORLD
View originalThis article was summarized and organized by BizCrush based on the original article from TECHWORLD. For exact quotations and full details, please refer to the original article.