Adaptive Dynamic Programming for Human Postural Balance Control

摘要

This paper provides a basis for studying human postural balance control about upright stance using adaptive dynamic programming (ADP) theory. Previous models of human sensorimotor control rely on a priori knowledge of system dynamics. Here, we provide an alternative framework based on the ADP theory. The main advantage of this new framework is that the system model is no longer required, and an adaptive optimal controller is obtained directly from input and state data. We apply this theory to simulate human balance behavior, and the obtained results are consistent with the experiment data presented in the past literature.