Demonstrator setup
In the project, a demonstrator consisting of two robot modules (Universal Robots UR5e) and a manual workstation will be set up to test and validate the proposed research topics on real industrial hardware. A ctrlX CORE is used as the controller within the modules, offering the possibility of expansion using self-developed applications. The aim is to demonstrate the use of robots in a flexible environment, using selected use cases from the areas of online trajectory planning, robot-robot-cooperation and human-robot-collaboration.
Human robot collaboration
For intuitive cooperation between robots and humans, reliable detection of the movement of a human body is necessary. The movements are registered with the help of a camera system as well as acceleration sensors and transferred to a combined model with the help of sensor fusion. Predicting human movement makes it possible for the robot system to avoid humans in time and support them in their work.
Publications:
- C. Cai, S. Liu, A Probabilistic Dynamic Movement Primitives Framework on Human Hand Motion Prediction for an Object Transfer Scenario, 2023. 22nd IFAC World Congress. LINK
Robot robot cooperation
In the project, an approach for online motion planning of robot manipulators in cooperation was developed. The algorithm enables robot systems to no longer be programmed "rigidly", but to take over motion planning independently and react to changes within their respective environment. The developed method was validated and optimized in the project using two scenarios. First, a scenario was developed in which a robot independently performs a pick & place task (sorting of different components) with the help of a vision system. In the second scenario, two robots cooperate in a shared workspace by jointly carrying out a disassembly process on several products. The robots exchange their current states at any time and can avoid each other with the help of the developed method.
Publications:
- N. Gafur, G. Kanagalingam, A. Wagner, M. Ruskowski, Dynamic Collision and Deadlock Avoidance for Multiple Robotic Manipulators, 2022. IEEE Access 10, 55766 - 55781. LINK
- N. Gafur, L. Weber, V. Yfantis, A. Wagner, M. Ruskowski, Dynamic path planning and reactive scheduling for a robotic manipulator using nonlinear model predictive control, 2022. 30th Mediterranean Conference on Control and Automation. LINK Experimental Results Video
Software architecture and standardization of interfaces
In addition to the autonomous movement of robots, the project also investigated the flexible expansion of the system with software and hardware components. For this purpose, possibilities for a plug & produce principle and individual control in the context of Industry 4.0 were considered. OPC UA was used to select a suitable communication protocol for modeling a semantically defined interface. Existing companion specifications were analysed, gaps identified and interfaces that meet all the requirements of a standardized interface were implemented for the modules using OPC UA. A capability-skill model was created to meet the need for flexible control and linking of the systems. The model was jointly consolidated in a working group of the Plattform I4.0. The expandability of the control systems used is made possible by an app-based approach. This enables the subsequent expansion of functionalities.
Publications: