作者: Takaaki Shiratori , Shunsuke Kudoh , Shin'ichiro Nakaoka , Katsushi Ikeuchi
DOI: 10.1109/IROS.2007.4399102
关键词: Engineering 、 Motion (physics) 、 Dance 、 Artificial intelligence 、 Motion generation 、 Scaling 、 Simulation 、 Computer vision 、 Upper body 、 Motion control 、 Frequency domain 、 Humanoid robot
摘要: This paper proposes a method to model the modification of upper body motion dance performance based on speed played music. When we observed structured performed at normal music playback and faster speed, found that detail each is slightly different while whole similar in both cases. phenomenon derived from fact dancers omit details perform essential part order follow To clarify this phenomenon, analyzed differences frequency domain, obtained two insights omission details: (1) High components are gradually attenuated depending musical (2) important stop motions preserved even when high attenuated. Based these insights, modeled our considering joint limitations humanoid robot has. We show effectiveness via some applications for generation.