ExSLeR: Development of a Robotic Arm for Human Skill Learning

The trend in robotics has shifted from collaboration with humans to learning and reproducing human skills, reflecting a growing social demand. In response, it is imperative to consider both hardware and software aspects in the design of robots. On the hardware side, the robot should be equipped with adequate sensors for mimicking human motion and force, and its design should meet necessary requirements such as workspace, degree of freedom, payload capacity, and weight, all of which are contingent upon the intended use of the robot. On the software side, the robot should be equipped with a real-time system and stable control algorithms to ensure safe operation. This paper presents the ExSLeR arm which meets the requirements for human skill learning. The performance of the ExSLeR arm is validated by a set of experiments through motion tracking with heavy payload and compliant interaction control tasks.

I. INTRODUCTION

The phenomenon of the interruption in the transmission of cultural heritage and expert knowledge has become a matter of concern in contemporary society [1], [2]. The potential of robots to acquire and reproduce the advanced skills of experts is seen as a means of mitigating these social challenges. In response to these social demands, robots have undergone a transformation from serving as cooperative aids to humans to becoming autonomous entities capable of imitating human skills [3]- [5].