Virtual Reality for Programming Cooperating Robots Based on Human Motion Mimicking

Abstract

While the new paradigm of integrating robots and humans in hybrid production systems is increasingly accepted by industry and research, there is still lack of methods for programming efficiently such robotic cells. Imitation learning is a user-friendly instruction technique to transfer new skills to robots. Despite the fact that the numerous research efforts have investigated the potential of generating robot motion by imitating the human behavior, there are still open research issues. This chapter elaborates on the concept of generating dual arm robot motion by imitating the human behavior in a virtual environment. The proposed method is implemented on a software tool via MATLAB. It is applied to an automotive industry case study, involving the pick-and-place of a cable within a repository. Open research issues regarding the human data capturing and learning algorithms, the human data modelling, as well as the classification and generation of robot motion are discussed.