基于约束物理系统和运动捕捉的人体运动仿真

为了能在基于PC平台的虚拟现实中实现更具真实感的人体运动仿真,该文提出一种改进的约束求解方法,将人体建模为基于约束关节树的骨骼结构。约束模型的求解采用基于线性互补问题的求解方法,同时结合相应的运动捕捉数据。实验结果表明,该方法所获得的运动数据不仅能满足交互式要求的帧速率,而且由于算法基于人体运动捕捉原始数据,使得虚拟人体与环境进行实时交互运动的过程更加真实。仿真结果以每帧的三维图像表示,可以形象地显示每秒60帧的人体运动。

A Physical based Human Motion Simulation Combined with Motion Capture Data

In order to make human motion simulation more realistic in Virtual Reality on PC platform, this paper introduces an improved method of modeling human body as a tree with constraint joints. The constraint human model is solved with a linear complementary problem solver. The motion capture data is also as the input. It makes the interaction of the virtual human body and the VR environment more believable. Simulation results are displayed as three dimension model with the rate of 60 frames per second. This expression style can intuitively display the result of human motion.