Vision-based grasp learning of an anthropomorphic hand-arm system in a synergy-based control framework

A reinforcement learning algorithm associates postural synergies to geometric object features for stable grasps of unknown objects.