A Model Predictive Approach to Control the Motion of a Virtual Player in the Mirror Game

Abstract

In this paper, we focus on the design of a feedback controller that drives a virtual player to follow or lead a human player in the mirror game. The movement of the end-effector of the virtual player is modeled by means of a feedback controlled Haken-Kelso-Bunz (HKB) oscillator or a damped harmonic oscillator, which is coupled with the observed motion of the human player measured in real time. A model predictive control algorithm is developed for the virtual player to generate humanlike trajectories while maintaining individual motor signature and guaranteeing bounded tracking error. Experimental results based on a prototype setup show the effectiveness of our strategy and its advantages over other existing algorithms.