인공지능기반무기체계의작전운용성능 (ROC) 요인에 관한 연구: 감시 및 정찰 분야를 중심으로 = A Study on the ROC(Required Operational Capability) Factors of Weapon Systems with Artificial Intelligence : Focusing on surveillance and reconnaissance

다국어 초록 (Multilingual Abstract)

A Study on the ROC(Required Operational Capability) Factors of Weapon Systems with Artificial Intelligence: Focusing on surveillance and reconnaissance Kang, Yeungtae Advisor : Prof. Jang, S-K, Ph. D. Department of Military Science The Graduate School of Chosun University In the recent war in Ukraine and the Israel-Hamas war, surveillance assets and strike assets applied with artificial intelligence were linked, resulting in remarkable criminal records. Therefore, the development of weapons systems to which artificial intelligence is applied, such as military intelligence and autonomous weapons in Korea, is more required than ever. In addition, the Republic of Korea is currently struggling with population cliffs due to low fertility rates and a lack of military service resources. As the total number of military enlistment resources has decreased, the criteria for enlisting have been relaxed, and a civic group called Senior Army has also appeared, claiming that they can become soldiers even though they are old. The military weapon system is also trying to design a troop-reducing weapon system in consideration of this lack of military service resources There are many voices that a weapon system to which artificial intelligence is applied should be developed, but there is a limitation that the ROC for making the necessary decision can not be materialized. Unfortunately, the desire and effort to apply artificial intelligence to the military weapon system have not yet led to the decision required by the Joint Chiefs of Staff. However, some rapid pilot projects are in progress to examine the possibility. As a rapid R&D project at the National Defense Advanced Acquisition Technology Institute, an integrated AI convergence mobile boundary system and a simulated flight training system for AI-based tactical development training are being promoted on a trial basis. However, some rapid pilot projects are underway to examine the possibility. As a rapid R&D project at the National Defense Advanced Acquisition Technology Institute, the AI convergence mobile integrated boundary system and the simulated flight training system for training AI-based tactical development are being promoted on a trial basis. In order to determine the need for a weapon system to which artificial intelligence is applied, it is necessary to discuss the question of what operational performance is required and which ROC. should be reviewed first. To support this, the purpose of this study is to identify the ROC. of a weapon system applying artificial intelligence and to derive priorities among the derived items. Research data on weapon systems to which artificial intelligence is applied are relatively limited. Therefore, in this study, after conducting literature research, it went through a process of being verified through interviews by a number of artificial intelligence and weapon system experts. After the verification process, the reliability and priorities of each ROC. item were analyzed through the AHP survey, a hierarchical decision-making method. The applied research method was conducted through ① literature research, ② interviews to verify it, and ③ AHP survey ④ results analysis, a hierarchical decision-making method. As a result of the study, three items were analyzed as the top items of the surveillance and reconnaissance weapon system to which artificial intelligence was applied: surveillance and reconnaissance performance, artificial intelligence performance, and payload performance. The results of the importance evaluation showed that despite the rapid development and potential of artificial intelligence, the surveillance and reconnaissance performance were relatively high, indicating that it was still treated with the most weight. After that, artificial intelligence performance and payload performance appeared in order. Therefore, even when examining a weapon system to which artificial intelligence is applied, it was found that the implementation of the ROC. of the weapon system, which is the most fundamental, is becoming more important. The ROC. of an evolutionary concept should be reviewed in order to apply artificial intelligence performance to operational performance. It may be challenging to anticipate the advancement of artificial intelligence-related cutting-edge technologies. Additionally, as the military demands, no decision should be made based on a single, high-performance ROC. from the outset. The initial ROC. should be determined in consideration of the current level of science and technology, and the improved ROC. should be determined in consideration of the advanced level of science and technology in the future. Through this study, we hope that research and discussions between the government, the military, academia, and developers who want to electrify weapons systems to which artificial intelligence is applied will be activated. Key words : Required Operational Capability, Artificial Intelligence, Weapon Systems, Surveillance, Reconnaissance

목차 (Table of Contents)

• 제1장 서론 1
• 제1절 연구 배경 및 목적 1
• 제2절 연구 범위 및 방법 4
• 제3절 선행연구 9
• 제2장 이론적배경 15
• 제1절 국내외 국방 인공지능 기술 개관 15
• 1. 인공지능 15
• 2. 인공지능 무기체계 개발동향 22
• 3. 인공지능 관련 법령 및 훈령 33
• 제2절 무기체계 획득 관련 이론 34
• 1. 국방기획관리체계 34
• 2. 소요제기 및 소요결정 37
• 3. 작전운용성능(ROC) 39
• 4. 획득관리체계 46
• 제3장 작전운용성능(ROC) 요인 도출 50
• 제1절 작전운용성능 요인 50
• 1. 감시정찰 성능 51
• 2. 인공지능 성능 59
• 3. 탑재체 성능 69
• 4. 소결론 72
• 제2절 전문가 델파이 결과 74
• 1. 전문가 집단 구성 74
• 2. 델파이 결과 76
• 제4장 인공지능무기체계작전운용성능우선순위분석 88
• 제1절 AHP 기법의 개요 88
• 제2절 AHP 계층구조 및 설문조사 91
• 1. AHP 계층구조 91
• 2. AHP 설문조사 91
• 제3절 우선순위 결과 분석 101
• 제5장 결 론 104
• 제1절 연구 결과 요약 104
• 제2절 정책적 시사점 106
• 제3절 연구의 한계와 향후 연구 방향 107
• 참고문헌 108
• 부 록 113