Telemanipulation of an industrial robotic arm using gesture recognition with Kinect

Abstract

Robot telemanipulation in Master-Slave mode is required for unstructured and hazardous environments as in case of high-radiation areas. The expensive master devices can be replaced by the cheaper Kinect camera and the telemanipulation can be done using gestures. The traditional motion mappings from master devices to the slave prove to be quite non-intuitive when used for gesture based telemanipulation. This paper proposes Maximum Velocity Drift Control approach for gesture based telemanipulation of an industrial robotic arm. The approach is quite intuitive for the operator and requires minimal amount of training. Since only the visual feedback is available in gesture based control, it becomes inadequate as in case of occluded views. Hence, to perform the safe telemanipulation, we have combined force-control with gesture based control for avoiding any collisions with the environment. The target tasks were Pick and Place and Liquid Handling operations which demonstrate position control and orientation control respectively. Further, various safety features are provided so as to have safe telemanipulation without any accidents.