Ontological Knowledge Management Framework for Grasping and Manipulation

A holistic task based grasping system requires the use of perception modules that are tied with knowledge representation systems in order to provide optimal solutions. Task based grasping has been a well-researched field. One of the most comprehensive systems for task based grasping is the KCoPMan system that uses semantic information in mapping and planning for grasping. However, this system as well as other lacks conceptual knowledge that can aid the perception module in indentifying the best objects in the field of view for task based manipulation. Furthermore, the system depends on full 3D models of objects in order to perform grasping. This restricts the scalability, extensibility, usability and versatility of the system. In this paper, we propose an alternate knowledge representation and inference framework based on the concept of functional and geometric part affordances in order to aid task based grasping. The performance of the system is evaluated based on complex scenes and indirect queries. KeywordsOntologies, Knowledge Representation, Grasp Affordances, ConceptNet