무인전투기의 최적 행위 도출을 위한 사례 기반 추론 기법

초록

Recently, technologies related to unmanned vehicles are getting attention, and research on automated vehicles are widely conducted. Although various algorithms have been proposed, their practical usage is limited as higher accuracy requires higher complexity. Particularly, it is important to infer decisions in a real-time manner for unmanned combat aerial vehicles (UCAV) as much as to obtain high accuracy of the decisions. Therefore, we propose a method for inferring optimal behavior of UCAV using case-based reasoning methods, where the optimal behavior is determined rapidly with comparable accuracy. Specifically, the proposed method infers the optimal behavior by searching similar situations with behaviors from database, where previous combat logs are stored in the form of cases. From the experiments using combat logs, the proposed method outperformed the compared methods in terms of both prediction time and accuracy. It is expected that the adoption of the proposed method to UCAV will contribute to the advances in the unmanned vehicle technology.

키워드