Animation of Human Walking in Virtual Environments

This paper presents an interactive hierarchical motion control system dedicated to the animation of human figure locomotion in virtual environments. As observed in gait experiments, controlling the trajectories of the feet during gait is a precise end-point control task. Inverse kinematics with optimal approaches are used to control the complex relationships between the motion of the body and the coordination of its legs. For each step, the simulation of the support leg is executed first, followed by the swing leg, which incorporates the position of the pelvis from the support leg. That is, the foot placement of the support leg serves as the kinematics constraint, while the position of the pelvis is defined through the evaluation of a controlcriteria optimization. Then, the swing leg movement is defined to satisfy two criteria in order: collision avoidance, and control-criteria optimization. Finally, animation attributes, such as controlling parameters and pre-processed motion modules, are applied to achieve a variety of personalities and walking styles.