Von: Marc Manz
Self-adaptive gripping devices and gripping reflexes to enhance robotic grasping
The manipulation of objects is a major ability for robotic systems for interaction with the environment. Mechanically self-adaptive mechanism pave the way for the lightweight design of grippers which perform reliable gripping for a wide range of objects and reduce the complexity for control and grasp planning. Weight restriction might appear as a potential problem regarding the mechanical design, but by learning from nature, control strategies become apparent that increase this ratio of payload-to-weight for dynamic cases. Humans vary the used force for grasping depending on the arising external forces and parameters like the friction between object and skin. The safety ratio of provided force and required force has to be minimized in order to reduce physical fatigue and avoid the damage of objects which have to be manipulated.
In this talk I speak about the development of a holistic concept for grasping which covers the mechanical design with inherent adaptive capabilities, the low level controller to reaction on unforeseen situations and the sensor system to measure or predict the current gripper state.