Grasping system that familiarizes itself with unfamiliar objects

Motivation
In FAMULA, researchers from the Cluster of Excellence Cognitive Interaction Technology (CITEC) are conducting basic research on self-learning robots. With this work, they are contributing to the future use of more complex, multi-fingered robot hands, which today largely remain too costly and too complicated for use in industry, for instance.  

Objective
The knowledge being gained from the large-scale project FAMULA can help service robots to familiarize themselves in new households or learn new work processes. To do this, the
robot has to create mechanical intelligence for these hands: It has to learn how everyday objects can be distinguished using visual cues (color, form) and tactile characteristics (weight, surface), as well as what matters when it comes to grasping or handling these objects.

Robot Hands
The robot hands mimic both the shape and agility of human hands. Their key feature are integrated tactile sensors, which not only allow the robot hands to grasp and handle objects like human hands, but also to touch and feel. If a glass slips away or an object is softer than initially assumed, this data is transmitted and the hands adjust the force of their grip. It is this sensitive and flexible grasping and the slip detection as well as dynamic grip force control that make for robust handling of everyday objects.

Floka, the Robot Head
Floka is a social robot with eyes, eyebrows, and a mouth in a cartoon-like appearance. It enhances the robot hands’ actions with facial expressions. Floka follows the movements of the
hands. Like this it creates a more natural interaction situation. Research at CITEC has shown that Floka is more readily accepted as an assistant and conversation partner by humans than a more technical-looking robot.

Participants
Neuroinformatics research group
Neurocognition and Action – Biomechanics research group

With Additional Participation from
Affective Neuropsychology research group
Applied Informatics research group
Cognitive Systems Engineering research group
Cognitive Neuroscience research group
Social Cognitive Systems research group
Semantic Computing research group
CITEC Central Lab Facilities

Contact

Prof. Dr. Helge Ritter
Bielefeld University
Cluster of Excellence Cognitive Interaction Technology (CITEC)
Telephone: 0521 106-12123
Email: helge@techfak.uni-bielefeld.de