Human motion control with physically plausible foot contact models

Abstract

The foot-to-ground contact model plays an important role in the simulation of highly dynamic motions, such as turns and kicks. In this paper, we propose a method for solving dynamically cumbersome slipping contact problems, which are frequently observed in highly dynamic motions. We employ and modify a combination of two different types of cones representing the inequality constraints of a contact model: the friction cone and the velocity cone. The friction cone makes character animation physically plausible while the velocity cone allows a character to perform a sharp turn without foot-to-ground penetration. Our system effectively simulates human behavior using an inverted pendulum on a cart (IPC) model and motion capture data. In the preprocessing step, we analyze motion capture data to extract meaningful information for the IPC model. At run-time, our system produces a physically simulated character by tracking the desired motion that is predicted by the IPC model. We formulate human motion control as a quadratic programming satisfying the proposed foot-to-ground contact constraints. Our examples show that the proposed system can produce physically plausible character animation without noticeable foot-to-ground contact artifacts.