http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
박성관(Sung Gwan Park),김동환(Dong Hwan Kim) 제어로봇시스템학회 2020 제어·로봇·시스템학회 논문지 Vol.26 No.11
In this study, the performance of autonomous flight was analyzed by introducing the PPO method as reinforcement learning for autonomous flight of drones. A simulator based on the dynamics of a drone was produced, and the performance of autonomous flight was confirmed when reinforcement learning was applied to a drone using this simulator. After that, the possibility of autonomous flight was confirmed by applying the PPO algorithm to the actual drone. Also, a lightweight embedded PC was attached to the drone to perform independent calculations to simultaneously construct obstacle avoidance and path planning.
협동로봇의 충돌 위험 지수 기반 심층 강화학습을 이용한 로봇의 안전 궤적 생성
박성관(Sung Gwan Park),김건우(Gunwoo Kim),신헌섭(Heonseop Shin),임성수(Sungsoo Rhim) 대한기계학회 2023 대한기계학회 춘추학술대회 Vol.2023 No.11
In general, while various safety methods are employed when using collaborative robots, human-robot collisions can still occur due to unforeseen accidents. As a result, human-robot interactions have typically been conducted at lower speeds to ensure safety. While this reduced speed makes the interaction safer for humans, it also slows down the robot, leading to decreased operational efficiency. In this paper, we introduce a reinforcement learning method that creates a trajectory designed to increase work speed without compromising safety which is based on the Collision Risk Index (CRI) indicator. Compared to traditional methods, our approach enhances safety within the same operation time.