Automated Commissioning of Offline-Generated Robot Programs

Abstract

The flexibility of industrial robots accounts for their importance in reconfigurable production systems. Advanced methods of offline programming based on 3D simulation promise seamless adaptation to changing circumstances. However, since digital planning models generally do not match reality, the commissioning of computer-generated robot programs involves extensive manual adjustment to the on-site conditions. This paper presents a concept for the automated commissioning of offline-generated robot programs to reduce downtimes of industrial robots and to increase the process consistency of offline robot programming. The presented approach aims for a hybrid motion control of industrial robots that automatically compensates for deviations between simulation and reality. For this purpose, robot poses are described not only by the joint positions but also by the anticipated sensor information. During process execution, robot poses are adjusted as required on the basis of the sensor data collected online.