Soundscape Generation for Virtual Human Robot Collaboration

Abstract

Despite increasing automation in manufacturing and assembly, production without humans is unlikely in the future. Instead, human-robot collaboration will play a central role with regard to future assembly processes. The accompanying changes in tasks require corresponding vocational training and continuing education measures for competence development. Training in a mixed reality (MR) learning environment, in which the handling of collaborating robots can be trained safely, promises great potential in this context. For a realistic learning experience, a soundscape is of high importance. Based on a virtual learning factory designed with Unity, this paper describes a prototypical integration of a realistic sound synthesis. This paper aims to generate a realistic sound from arbitrary positions depending on the current position, load and speed during movement. For this purpose, a characteristic sound is synthesized for each robot joint and adapted in volume and pitch to the current operating conditions such as rotational speed and acting torque at runtime. This allows the user to estimate the robot movement without looking at it. For a training within collaborative assembly operations, this experience is of particular importance in order to create acceptance of robots within the same workspace. The results are validated in a practical investigation.