A control allocation approach to haptic control of underwater robots

This paper presents a new set of approaches for teleoperation of remotely operated vehicles (ROVs), or autonomous underwater vehicles (AUVs) with on-board cameras. We address the control allocation problem for underwater vehicles, i.e., how to best distribute the control forces so that the robot behaves as intuitively as possible based on the operator's commands. This calls for intelligent control allocation in order to obtain intuitive control of the camera as we do not control the camera directly, but rather the underwater robot, with slow and inaccurate dynamics, in addition to normally just a few degrees of freedom to the camera, which has a lot faster dynamics. For the presented approach, the operator only needs to control the camera motion, while the vehicle will follow the camera naturally. The operator is thus able to control a system that is kinematically very different from the haptic device very intuitively. The control allocation problem, i.e., how the vehicle and camera actuation is applied, is then performed automatically so that the operator can concentrate on the camera motion.