이병선 안상현 이기한 김동욱 전우성 박상철 2023-06 2508-4003 https://aurora.ajou.ac.kr/handle/2018.oak/37834 https://www.kci.go.kr/kciportal/ci/sereArticleSearch/ciSereArtiView.kci?sereArticleSearchBean.artiId=ART002963516 This study presents a novel approach to define air-to-air combat patterns and scenarios based on basic fighter maneuvers with a focus on realism. To accomplish this goal, we developed a real time air-to-air combat simulation environment for personnel training, which is free from affilia tion or faction constraints. The simulation environment is designed to concentrate on 1-on-1 air to-air combat simulations based on gun engagements. In order to facilitate swift decision-mak ing during aerial combat, we developed an KNN based risk assessment model. This model enables accurate and timely judgment of air-to-air combat situations. By applying the risk assessment model to the simulation, we effectively assessed the situation of friendly aircraft and conducted appropriate maneuvers, resulting in increased attack and evasion success rates. Our findings demonstrate significant benefits in military training and operations simulations through air-to-air combat simulations using the risk assessment model. This research provides a founda tion for future development of advanced combat training and mission planning systems Kor 한국CDE학회 위험도 평가 모델을 활용한 훈련 목적 공대공 전투 시뮬레이션 Training-oriented Air-to-air Combat Simulation Using a Risk Assessment Model Article 117 2 107 한국CDE학회 논문집 28 한국CDE학회 논문집, Vol.28 No.2, pp.107-117 10.7315/CDE.2023.107 Air-to-air combat Basic fighter maneuvers Machine Learning Risk assessment Article