In: Zhigeng Pan, Ruth Aylett, Holger Diener, Xiaogang Jin, Stefan Göbel, Li Li (Eds.): Lecture Notes in Computer Science, 3942: Technologies for E-Learning and Digital Entertainment: First International Conference, Edutainment 2006, Hangzhou, China, April 16-19, 2006, pages 1027-1036. Springer-Verlag, March 2006. URL: http://www.springerlink.com/openurl.asp?genre=article&id=doi:10.1007/11736639125,.
Abstract: This paper presents a set of technologies to generate human animations based on captured motion data. Firstly, original coarse motion data need to be normalized as standard data set. Secondly, spatial and temporal characteristics of human locomotion are calculated according to experience equations, anthropometric statistics data and user-defined interface parameters. Trajectory of human's root site is scaled and rotated to fit real walking distance and direction. Then, a simulation structure based on place/transition Petri Net is designed and conducted to start, run, pause and end simulation loop. During the update phase of the loop, position and posture of virtual figure in each frame are generated through interpolating from scaled motion data.