Reactive human-robot interaction learning in virtual environments

Abstract

The study of human-robot interaction (HRI) is of considerable interest today in the field of robot technology. Current methods for HRI require interactive learning, which can be slow, computationally demanding, complex, and also unsafe depending on the nature of the robot To solve these problems, this paper proposes a method of interaction learning in a virtual environment where a virtual robot can learn to interact with a virtual human that is designed to mimic human movements through imitation learning. Then the result of this virtual robot can be applied to the real robot after the interaction learning is completed. ©2014 International Information Institute.