Prototyping proactive and adaptive techniques for human-robot collaboration in manufacturing using virtual reality

Abstract Human-Robot Interaction (HRI) has emerged in recent years as a need for common collaborative execution of manufacturing tasks. This work examines two types of techniques of safe collaboration that do not interrupt the flow of collaboration as far as possible, namely proactive and adaptive. The former are materialised using audio and visual cognitive aids, which the user receives as dynamic stimuli in real time during collaboration, and are aimed at information enrichment of the latter. Adaptive techniques investigated refer to the robot; according to the first one of them the robot decelerates when a forthcoming contact with the user is traced, whilst according to the second one the robot retracts and moves to the final destination via a modified, safe trajectory, so as to avoid the human. The effectiveness as well as the activation criteria of the above techniques are investigated in order to avoid possible pointless or premature activation. Such investigation was implemented in a prototype highly interactive and immersive Virtual Environment (VE), in the framework of H-R collaborative hand lay-up process of carbon fabric in an industrial workcell. User tests were conducted, in which both subjective metrics of user satisfaction and performance metrics of the collaboration (task completion duration, robot mean velocity, number of detected human-robot collisions etc.) After statistical processing, results do verify the effectiveness of safe collaboration techniques as well as their acceptability by the user, showing that collaboration performance is affected to a different extent.