Dynamic Simulation of a Complex Human Figure Model with Low Level Behavior Control

Abstract

Advances in computer hardware and software technology allow the simulation of natural phenomena in increasing levels of complexity. This research is concerned with simulating the articulated movements of humans using the laws of physical motion, and contributes to the fields of computer animation and biomechanics. A 90 degree of freedom model of a human figure is developed, and an efficient dynamic simulator is employed to create and analyze physically based, computer-generated motions. The foot of the simulated human figure has been modeled with a significant amount of kinematic complexity, with 28 degrees of freedom per foot. A joint-level control layer uses springs and dampers to control postures and movements. Inverse dynamics and inverse control can be used to calibrate the spring actuators to exactly achieve specified postural goals. A framework for higher level control is implemented, although specific tasks require tailored control strategies. Several example tasks are simulated and described, including a stable standing posture and the stepping phase of walking using passive dynamic effects to generate much of the motion. Together, the simulator and biomechanical model create a framework that can be used to address problems in computer animation and biomechanical research, and eventually, in a clinical setting, to assist doctors in analyzing the problems of specific patients.