Skill-based programming of complex robotic assembly tasks for industrial application

• Published in: Elektrotechnik und Informationstechnik Vol. 136; no. 7; pp. 326 - 333
• Main Authors: Akkaladevi, Sharath Chandra, Pichler, Andreas, Plasch, Matthias, Ikeda, Markus, Hofmann, Michael
• Format: Journal Article
• Language: English
• Published: Vienna Springer Vienna 01-11-2019; Springer Nature B.V
• Subjects: Assembly; Automation; Computer Hardware; Domains; Electrical Engineering; Engineering; Graphical user interface; Industrial applications; Knowledge acquisition; Mass production; Order parameters; Originalarbeit; Programming; Robotics; Robots; Software Engineering/Programming and Operating Systems; Task complexity; Workflow; Mensch-Roboter Kooperation; human-robot cooperation; skill-based learning; Skill-basiertes Lernen
• ISSN: 0932-383X; 1613-7620
• DOI: 10.1007/s00502-019-00741-4

In recent years, a paradigm shift is underway as robots leave their typical application field and move into domains that have been untouched by robotic automation. These new kinds of automation systems allow more product variations, smaller life cycles, smaller batch sizes and pave the way from mass production to mass customization. This is due to completely new breed of safe robot technology but also novel ways of setting up new applications like e.g. kinesthetic programming. However, the topic of reducing the programming effort for complex tasks using natural modes of communication is still open. This paper addresses the key developments in this field, shows different ways of programming, and gives relevant use cases in industrial assembly. The technology coverage starts with an online workflow editor called XROB that allows easy-to-use setup of process workflows and related skill parameters. However, in order to reduce the programming effort, a novel way to demonstrate process trajectories by using instrumented hand guided process tools is presented. Finally, the paper gives an overview of a promising approach that allows programming without touching the robot just by demonstrating the process by an expert. The semantic relations between activities executed by the human and robot skills are captured to learn the task sequence of the assembly process. The acquired process knowledge is refined to execute robotic tasks with the help of an interactive graphical user interface (GUI). The system queries the user for feedback, asking for specific information to help the robot complete the task at hand. The given examples show the usability of flexible programming tools in the automation chain and the presented results provide strong evidence of the technological potential in the field.