Voxel Map Based Collision Detection For Underwater Manipulators

This paper presents a real-time collision detection algorithm for underwater robotic manipulation. It is capable of detecting collisions between multiple manipulators and between manipulators and obstacles in the workspace. Assuming operations are typically carried out on artificial structures with a known geometry, it is possible to model the environment from a priori known CAD models. The position of objects in the workspace is tracked using a vision based approach that relies on fiducial marker detection. The proposed collision detection method is developed, integrated into a commercial ROV manipulator control system, and successfully evaluated in simulations and experimental setup using a real industry standard underwater manipulator.