Human-machine interaction (HMI) technologies enable the automation of various manufacturing and assembly applications while maintaining high flexibility. In this context, human-robot collaboration (HRC) capable robots should no longer function as autonomous systems, but much more as assistance systems or as colleagues for workers. In connection with shorter product life cycles, increasing variant diversity and individualization, the challenge arises to set up flexible robot systems, which can be reprogrammed and commissioned with little effort in a short period of time with preservation of the required accuracies [1].
Therefore, intelligent path planning is essential for development of flexible robot systems. In this paper the development of different approaches are presented that allow the worker on the shop floor to rapidly and easily program a robot to implement new motion tasks based on a camera and sensor system without programming knowledge. Thereby various points are selected manually on an image. Then point-controlled processes or even continuous paths can thus be specified and transferred to the robot. A more automated approach based on further image processing methods allows contour detection and finally the generation of process-specific paths. The approaches are not only limited to simple geometries and flat surfaces, but also handle free-form surfaces. To achieve a high flexibility, the setup envisages mounting a 3D camera on the end effector of the robot.
The above mentioned approaches will be conceptualized on sealing and riveting applications, which are common in the aircraft production. These application examples have a high demand for flexible robot systems, since CAD data for robot path planning is often not available for aircraft assembly and therefore robot programming has to be done on site. Furthermore, aircraft production is characterized by a high level of manual activities and the dimensions in the aircraft industry require the permanent re-localization and re-commissioning of the robot system. The complete implementation and validation of the presented approaches and use cases will be covered in the frame of future works.