Soft-Grasping With an Anthropomorphic Robotic Hand Using Spiking Neurons

Evolution gave humans advanced grasping capabilities combining an adaptive hand with efficient control. Grasping motions can quickly be adapted if the object moves or deforms. Soft-grasping anthropomorphic is a great capability for robots interacting objects shaped humans. Nevertheless, most robotic applications use vacuum, 2-finger custom made grippers. We present biologically inspired spiking neural network (SNN) soft-grasping to control hand. Two loops are combined, one from motor primitives and compliant controller activated by reflex. The finger represent synergies between joints different affordances. Contact detected mechanism based on inter-neuron circuits in spinal cord trigger reflexes. A Schunk SVH 5-finger was used grasp shapes, stiffness sizes. SNN without knowing exact properties of objects. online learning proved sensitive, allowing even balloons. In contrast deep approaches, our requires example each motion train primitives. Computation inverse kinematics complex contact point planning not required. This approach simplifies providing similar features as human physical imitation biological system implemented completely provide new insights into mechanisms.