Intuitive robot teaching by hand guided demonstration

Abstract

In this paper, we have developed an intuitive robot teaching method based on hand guided demonstration. A depth camera is used to recognize various hand gestures which are used as the commands to control the robot gripper. Support Vector Machine (SVM) technique is used to classify different hand gestures. By an inverse kinematics approach, the human hand and robot endpoint workspace mapping has been performed. A preliminary experiment has been carried out on the Baxter robot simulator to verify the effectiveness of the proposed method. Test results indicates that the proposed hand guided robot teaching is intuitive, natural and effective and is of great potential for applications in human robot interaction.