S1000D 기반 IETM과 언어모델을 활용한 무기체계 정비도 개선 연구

이동균^* ; 김제용 ; 황희원 ; 이계신

LIG넥스원 미사일시스템 IPS연구소 수석연구원

Improving Weapon System Maintainability with S1000D-based IETM and Language Model

Dong-Kyoon Lee^* ; Je-yong Kim ; Heewon Hwang ; Kye-shin Lee

Chief research engineer, Missile Systems Integrated Product Support R&D Lab, LIG Nex1

Correspondence to: *Dong-Kyoon Lee Missile Systems Integrated Product Support R&D Lab, LIG Nex1 20, Daewangpangyo-ro 851 beon-gil, Sujeong-gu, Seongnam-si, Gyeonggi-do, Republic of Korea Tel: +82-31-326-9257 E-mail: dongkyoon.lee@lignex1.com

초록

본 연구는 S1000D 기반 IETM과 언어모델 챗봇을 결합하여 무기체계 정비도 향상 방안을 제안한다. 기존 정비는 기술교범과 정비요원 숙련도에 크게 의존하여 결과의 일관성이 부족했다. 이를 개선하기 위해 S1000D 데이터 모듈과 RAG, 프롬프트 엔지니어링 기법을 적용한 정비지원 프로토타입을 설계하였다. 다양한 언어모델을 비교·검증한 결과, 제안된 시스템은 단순 검색보다 절차 이해와 효율성이 높아 고장탐지 및 배제 과정에서 실질적 지원이 가능함을 확인하였다. 본 연구는 군 특성에 적합한 소형 언어모델 최적화와 데이터 모듈 확장을 통한 정비지원체계 발전 가능성을 제시한다.

Abstract

This study presents an approach to improve weapon system maintainability by integrating S1000D-based IETM with a language model chatbot. Conventional maintenance relies heavily on manuals and personnel expertise, often leading to inconsistent results. To address this, a prototype system was designed using data modules, RAG and prompt engineering. Comparative evaluation of multiple language models showed that the proposed system enhances procedural clarity and efficiency beyond simple search, effectively supporting fault detection and isolation. The findings highlight the potential of compact offline language models and extended S1000D modules for future military maintenance support systems.

Keywords:

IETM(Interactive Electronic Technical Manual), RAG(Retrieval-augmented Generation), Chatbot, Fault Detection and Isolation, Missile Maintenance, Language Model

키워드:

챗봇, 고장탐지 및 배제, 유도탄 정비, 언어모델

References

AeroSpace and Defence Industries Association of Europe, “S1000D - International specification for technical publications using a common source database.”
Joint Air Power Competence Centre, “How Large Language Models are Transforming Modern Warfare.” JAPCC Journal, Edition 37, pp. 68-76, 2024.
Dong-Kyoon Lee et al. “Maintainability Prediction of Guided Missile Based on Machine Learning Using Field Data,” Journal of the Korea Academia-Industrial cooperation Society, Vol. 23, No. 2, pp. 518-526, 2022.
Kyung Han Lim et al. “Korea’s Artificial Intelligence Defense Policy Development Plan: Focusing on Cases of US and Japanese AI Defense Policies,” Journal of KNST, Vol. 7, No. 4, pp. 586-590, 2024.
Gwan-Woo Goh et al. “Mil-BART: Military Domain Specialized Pre-trained Language Model Using Defense Standard Natural Language Dataset,” Journal of KIIT, Vol. 22, No. 1, pp. 51-59, Jan. 31, 2024
Jang Wan Hur, “Analysis of Inspection Interval According to the Operating Platform of Overseas Purchase Missile,” Journal of Aerospace System Engineering, Vol. 15, No. 1, pp. 63-71, 2021.
D. J. Theunissen and R. Owen Holbrook, “Effects of Environment and Aging upon Missile Reliability,” Annual Reliability and Maintainability Symposium, 1998.
Gao, Y. et al. “Retrieval-augmented Generation for Large Language Models: A Survey,” arXiv preprint arXiv:2312.10997, , 2023.
Ji, Z. et al. “Survey of Hallucination in Natural Language Generation,” arXiv preprint arXiv:2303.12013, , 2023.
Lewis, P. et al. “Retrieval-augmented Generation for Knowledge-intensive NLP Tasks”. arXiv preprint arXiv:2005.11401, , 2020.