Development of a Graph-Based Translation From BPMN to Executable Sequences for Industrial Robotic Systems

Abstract

The demands regarding high mix, low volume manufacturing and faster product life cycles make flexible production indispensable. Collaborative robots are widely regarded as an enabler for this flexible production. Further, they also achieve the goal of human-centered production envisioned by Industry 5.0. However, its installation requires significant efforts by skilled specialists and robotics experts for robot programming. In order to improve accessibility for shop-floor workers, the focus in previous works lies on the combination of graphical/function-based and declarative programming that supports fast reconfiguration. The business process model notation (BPMN) was used for the user input of hardware-independent robot skills. Further, the so-called SAMY-Core was developed to generate control commands for the respective hardware. Based on these two components, this work focuses on the concluding translation of BPMNs to machine executable processes as the final component within the SAMY framework to finalize its entire pipeline from the user input to the hardware-specific code execution. For the translation, the SAMY-BPMN is processed to a graph, which contains all skills and can return the correct following actions by means of states of the robot system. As a result, it is shown that this translation and thus the entire pipeline is feasible, allowing non-expert users to change the system both quickly and easily.