Nowadays, before industrial robots can effectively be used in production environments, the robots undergo a time-consuming teach-in process. As a consequence, repeatedly conceptualizing the three-dimensional movement options for a robot, mapped in the form of a coordinate system, and programming the movement paths represents an extremely complex task and restricts the flexibility.
We focuse in the AUTOWARE project on the development of Visual Computing algorithms that will give robots cognitive abilities and increase their flexibility when it comes to perform unexpected tasks in cooperative scenarios. Hence, complex teach-ins will not be necessary anymore. These developments will be attractive for customers (especially automotive and aerospace industry). Regarding the transformation of the cooperative assembly scenarios we address the following research aspects:
- 3D-reconstruction of the working environment and segmentation of point clouds
- machine learning
- image processing and object recognition
- interaction concepts and and visualization von intention
- cognitive robotics