| Title |
Exploring the Feasibility of Local Small-scale Large Language Models for OBTS of ECS |
| DOI |
https://doi.org/10.5573/ieie.2026.63.7.135 |
| Keywords |
Large language model; Engineering control system; On board training system; AI integration |
| Abstract |
This paper investigates the application of a locally deployed, small?scale large language model (LLM) to the On?Board Training System (OBTS) within an Engineering Control System (ECS). As generative?AI technologies advance, services based on LLMs have emerged, prompting research on their adoption in the defense sector. In naval combat systems, AI integration is proceeding incrementally, starting with target?monitoring and data?cleaning functions, while large?ship environments are exploring LLM?based command?decision?support systems. In contrast, research on AI adoption for the ECS is still limited. To address this gap, we examine the feasibility of introducing AI into the ECS by focusing on the OBTS, which exerts the smallest impact on overall system architecture. Considering the characteristics of OBTS, we identify technical requirements, assess system?integration possibilities, and propose utilization scenarios within the OBTS. Based on these analyses, the overall possibility of system implementation is assessed. By demonstrating the feasibility of rapidly integrating and utilizing local small LLM, the findings of this study are expected to contribute to the adoption of AI technology within the field of ECS. |