Collaborative lightweight robots are a trend in industry. They are comparatively cheap. Developments in the field of machine learning make them increasingly flexible and easier to use. The challenges for research are obvious. The machines must be equipped with cognitive intelligence in order to adapt to a changed environment and understand new tasks. The aims of the project go far beyond the state of the art of research.
The focus is on “transfer learning”: from manual human activities to the robot and from a product or process variant to a similar one.
The main research objectives of the project are
1. mapping of human motion to the robot
2. the “understanding” of temporal task correlations and process parameters by the robot
3. adaptability to similar processes with as few new examples as possible
No external cooperation is envisaged for the implementation of the project.
Project name:
LERN4MRK: Modellieren, Erlernen und Abstrahieren von Prozessen für die Mensch-Roboter Kooperation
Funding:
bmvit
Duration:
01.07.2017 – 30.06.2021
Publikationen
S.C. Akkaladevi, M. Plasch, and A. Pichler, “Skill-based learning of an assembly process” Elektrotech. Inftech. (2017) 134: 312, Springer Vienna. https://doi.org/10.1007/s00502-017-0514-2
C. Heindl, T. Poenitz, G. Stuebl, A. Pichler, and J. Scharinger, “Spatio-thermal depth correction of RGB-D sensors based on Gaussian Processes in real-time” in The 10th International Conference on Machine Vision, to be published, 2017.
