Tele-assistance task planning for dextrous underwater manipulation

An approach to operator interaction for a three-fingered dextrous manipulator is described, using a tele-assistance method termed blind grasping. Starting with approximate co-ordinates for the object location, a series of finger contact movements is used to establish the object geometry and candidate stable grasp locations. The approach is implemented within the framework of the preference planning architecture, a reactive tele-assistance task planner that uses geometric and semantic world models to trigger a hierarchical set of task sequences. For the first prototype AMADEUS underwater gripper, we demonstrate that variously shaped objects centred between the open gripper fingers can generally be grasped using simple closure. However, off centre displacements as small as 1 cm require multiple touch samples to acquire the geometry with sufficient accuracy to choose stable grasp points