A visually guided robot and a neural network join to grasp slanted objects

Abstract

In this paper we introduce a method for model-free monocular visual guidance of a robot arm. The robot arm, with a single camera in its end-effector, should be positioned above a target, with a changing pan and tilt, which is placed against a textured background. It is shown that a trajectory can be planned in visual space by using components of the optic flow, and this trajectory can be translated to joint torques by a self-learning neural network. No model of the robot, camera, or environment is used. The method reaches a high grasping accuracy after only a few trials.