Enhancing the familiarity for humanoid robot pepper by adopting customizable motion

Abstract

Most of the works use sensor to recognize human gesture and teleoperate robot in real time for different usage. One of the purpose is social learning which human learn behavior and take place in the social context. Our research aims to let user create customizable gesture and motion for proper condition. This paper presented the method to replicates human gesture for robot imitation. If robot can adapt human like behavior to interact with human, it is potentially helpful to build intimate between human and robot. Human gesture were recorded and unnecessary noise was removed to smooth robot's motion. Furthermore, by changing motion speed it is possible to create different feeling so that it might be arouse human's interest and spice up the interaction. The experiment results show the parameter pair to implement motion on robot.