Robot Programming for Manipulators through Volume Sweeping and Augmented Reality

Today's industrial robots require that human operators teach motions in advance. However, conventional methods for robot programming need deep knowledge and skills about robots or great effort for inputting information of working environment into computers. Therefore, a robot programming method in which everyone can easily teach robots "good" motions is demanded. For this purpose, our group proposed a robot programming method that uses manual volume sweeping by operators and automatic motion planning together to generate motion plans with short cycle times. Because a swept volume is the space through which the robot has passed without collision, it is movable space of the robot that can be used in motion planning. In this paper, we proposed using augmented reality in this programming method. We constructed a system in which operators can perceive obtained swept volumes and generated paths intuitively through augmented reality. Teaching experiments showed that non-skilled operators can make a robot move in shorter time than teaching/playback by direct teaching.