Within recent 30 years, we have witnessed the progress of dexterous robotic hands from its designing and controlling aspects. However, how to use these advanced robot hands to implement skillful tasks like human are still very challengeable. Our expectations on the use of the robotic hands are rather high, the use cases in the robotic application are very few. We are considering the following three challenges for available robotic hands:
· It is still difficult to robustly, adaptively and dexterously control a robotic hand given the high mechanical complexity of the devices
· It is not clear how to implement complex tasks using the robotic hand equipped with limited sensing capability in the unstructured environment
· It is still an open question to represent and transfer human’s manipulation skills to a robotic hand which has a similar configuration space
In order to solve the aforementioned challenges, one solution will be merging the newest findings in neuroscience, cognitive science, machine representation, and learning domains, e.g. how human are using their hands skillfully and how humans are using their hands as an important recognition tool to explore and learn the unknown world. We believe this can lead to the real “manual intelligence” which not only can largely improve dexterous control of the robotic hand but also exploiting the robotic hand’s action-perception loop to autonomously understand the unstructured environment.
Within this workshop we will bring together experts from the different domains, e.g. the developers of dexterous hands, control scientists focusing on grasping, planning and computer scientists studying machine learning to discuss progress and challenges of hand’s dexterous manipulation, foster potential collaborations, and reinforce the strict link among such interdisciplinary research fields to facilitate progress in this community.
Central to the discussion will be three key questions:
· How to integrate multi-modal sensing for improving the robotic hand’s dexterous capability autonomously
· How to represent human’s manual skills and transfer them to robotic hands?
· How to define the benchmark to evaluate the dexterous capability of robotic hands?
The workshop topics include but are not limited in the following
· New type of dexterous robotic hand
· Human’s manual skill representation
· Robotic hand’s grasping planning and control
· Unknown objects in-hand manipulation
· Sensory-based robotic hand grasping and manipulation
· Neuro-inspired control for grasping and manipulation
· Machine learning techniques for grasping and manipulation
Prof. Tamim Asfour, Karlsruhe Institute of Technology, Germany
Prof. Shaowei Fan, Harbin Institute of Technology, China
Prof. Huaping Liu, Tsinghua University, China
Prof. Abderrahmane Kheddar, CNRS-UM LIRMM, France
Prof. Tetsuya Ogata, Waseda University, Japan
Prof. Weiwei Wan, Osaka University, Japan
Dr. Qiang Li, Bielefeld University, Germany
Dr. Maximo Roa, German Aerospace Center, Germany
Dr. Filipe Veiga, Technische University Darmstadt, Germany
Submission about the latest results is also welcomed! Prospective participants are required to submit a 2-page abstract. Accepted contributions will be presented during the workshop as posters. Submissions must be sent in PDF format, following the IEEE conference style (two-columns), to:
indicating [Humanoids 2018 Workshop] in the e-mail subject.
Workshop day: November 6
Dr. Qiang Li, Neuroinformatics Group / CITEC, Bielefeld University, Germany
Dr. Zhaopeng Chen, Robotics and Mechatronics Center, German Aerospace Center (DLR), Germany
Dr. Junpei Zhong, AI Research Center, National Institute of Advanced Industrial Science and Technology (AIST), Japan
Prof. Chenguang Yang, College of Engineering, Swansea University, UK
Prof. Helge Ritter, Neuroinformatics Group / CITEC, Bielefeld University, Germany