Data based steering of virtual human using a velocity-space approach
Résumé
In this paper, we analyze the habits of human walk. By plot- ting and studying some real experimental motion-captured data, we es- timate the bounds of the velocity and acceleration of the pedestrian and achieve the admissible motion set. We introduce the velocity-space which means the set of the whole admissible velocities for one step in the pro- cession of human walk. In addition, a model describing the dependence of the next movement on the previous one is designed. It provides an interior constraint for pedestrian in steering model. With two examples, we show the adaptability and efficiency of the model.