http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
- 中文 을 입력하시려면 zhongwen을 입력하시고 space를누르시면됩니다.
- 北京 을 입력하시려면 beijing을 입력하시고 space를 누르시면 됩니다.
RISS 인기검색어
- 검색키워드
- 저자 : Yunfei Ai (검색결과 1 건)
- 무료
- 기관 내 무료
- 유료
-
Learning Multiple-Gait Quadrupedal Locomotion via Hierarchical Reinforcement Learning
Lang Wei,Yunxiang Li,Yunfei Ai,Yuze Wu,Hao Xu,Wei Wang,Guoming Hu 한국정밀공학회 2023 International Journal of Precision Engineering and Vol.24 No.9
Over long periods of evolution, legged animals have developed the capability to use a variety of gaits to move efficiently and flexibly at different speeds. To enable quadruped robots to acquire this ability, this study proposes a two-stage training hierarchical framework that can have quadruped robots generate energy-efficient multiple-gait locomotion, consisting of a gait selection policy module and a react controller module. The parameters of both modules are optimized using reinforcement learning. The experimental results in the simulation demonstrate that the proposed method can generate energy-efficient multiple-gait quadrupedal locomotion compared to previous methods. To validate the robustness and effectiveness of the method, we constructed a closed-chain quadruped robot and deployed the controller trained by the method to the robot. The experimental results in the real world suggest that the controller can enable the robot to move stably and efficiently in different gaits. The main contribution of this paper is that the authors propose a novel hierarchical framework, which makes quadruped robots use an optimal gait at a specific speed and smoothly switch to another one after getting a different speed command. These behaviors are automatically produced through simulation training, eliminating the need for the tedious work of designing gaits and modulating controllers. Experimental results showcase that the proposed method has significant advantages compared to previous methods.
연관 검색어 추천
이 검색어로 많이 본 자료
활용도 높은 자료
