Combining Deep Learning for Visuomotor Coordination with Object Identification to Realize a High-level Interface for Robot Object-picking

IEEE-RAS International Conference on Humanoid Robots (Humanoids), pages 612--617, doi:10.1109/HUMANOIDS.2017.8246935 - Nov 2017.
Associated documents : combining-deep-learning-preprint.pdf [1.8Mo]   http://dx.doi.org/10.1109/HUMANOIDS.2017.8246935
We present a proof of concept to show how a deep network for end-to-end visuomotor learning to grasp is coupled with an attention focus mechanism for state-of-theart object detection with convolutional neural networks. The cognitively motivated integration of both methods in a single robotic system allows us to realize a high-level interface to use the visuomotor network in environments with several objects, which otherwise would only be usable in environments with a single object. The resulting system is deployed on a humanoid robot, and we perform several real-world grasping experiments that demonstrate the feasibility of our approach.

 

@InProceedings\{EKGNW17,
  author       = "Eppe, Manfred and Kerzel, Matthias and Griffiths, Sascha and Ng, Hwei Geok and Wermter, Stefan",
  title        = "Combining Deep Learning for Visuomotor Coordination with Object Identification to Realize a High-level Interface for Robot Object-picking",
  booktitle    = "IEEE-RAS International Conference on Humanoid Robots (Humanoids)",
  pages        = "612--617",
  month        = "Nov",
  year         = "2017",
  doi          = "10.1109/HUMANOIDS.2017.8246935",
  url          = "https://www2.informatik.uni-hamburg.de/wtm/publications/2017/EKGNW17/combining-deep-learning-preprint.pdf"
}

» Manfred Eppe
» Matthias Kerzel
» Sascha Griffiths
» Hwei Geok Ng
» Stefan Wermter