Universität Stuttgart
Permanent URI for this communityhttps://elib.uni-stuttgart.de/handle/11682/1
Browse
1 results
Search Results
Item Open Access Online motion planning for dual arm industrial robots(Stuttgart : Fraunhofer Verlag, 2020) Beuke, Felix; Verl, Alexander (Prof. Dr.-Ing. Dr. h.c. mult.)The trend toward consumer products of ever-increasing individuality poses several challenges for the domain of manufacturing automation. One of these challenges relates to the flexibility of devices used in industrial automation. In classical low-variability and high-throughput automation, plants could be designed and financed with several years of constant uptime in mind. In contrast, automated manufacturing of more and more individualized goods requires adapted manufacturing processes with shorter cycles and more flexible automation equipment. The higher variability and shorter life-cycle of manufactured goods often entails an uncertainty that makes investment in highly-specialized but inflexible automation equipment unprofitable. This creates a growing need for more flexible equipment that can be reused when the produced goods or parts of the production process change. Against this backdrop, the field of collaborative robotics has received increased interest and development activity over the past years. These robots can work in close proximity to humans and do not need to be fenced off in dedicated safety areas, thus enabling a more flexible use of the equipment. The idea of flexibility is taken further in this context by the development of dual-arm robots that mimic the morphology of the human upper body with two arms mounted on one common body. While these devices potentially offer increased flexibility and can work on a set of tasks otherwise reserved for humans (e.g., dual-arm handling, parallel work with both arms) the close proximity of the arms brings about new problems in the application of these robots. Using available programming methods for creating collision-free and coordinated motions for dual-arm robots is often complex and leads to long application times, which takes away a considerable part of the added flexibility introduced by dual-arm robots. To enable users to leverage the potential flexibility of industrial dual-arm robots more easily, this work develops a coordinated motion planner and associated control infrastructure for dual-arm industrial robots. Together, these components relieve the programmer of the responsibility to prevent arm collisions by specifying an exact temporal synchronization for both arms. Instead, collision-free and coordinated motions are automatically planned and executed based on a geometrical model. Moreover, it can serve as an application basis for more high-level planners for dual-arm industrial robots. The proposed planner and control architecture are implemented and evaluated on a real dual-arm robot.