Kinematisch hyperredundante Greifersysteme mit neuronaler Steuerung

Conventional robot grippers show a limited performance in case that the objects are soft or variably structured. Objects difficult to handle are, e. g. liquid filled bags or amorphous goods. Additional issues arise if the object position is weakly defined in the working-space, as it may be the case for micro- or nano-robotic applications. The following article introduces the concept of a two-dimensional tentacle-shaped gripper with distance sensors which accomplishes an automatic approach coordination and grasping sequence for the handling of objects, coordinated by a neural controller.