An OWL Ontology for Local Situation Awareness based on Object Hierarchy Propagation
Frank Dittrich, Heinz Woern · International Symposium on Robotics · 2014
In our paper we present a generic modeling and inference scheme for spatially resolved reasoning about situations, activities and states of distributed objects in a 3D scene, using an ontology based on the Web Ontology Language. As a basis for classification and local object related inference, assertions about the object and assigned activities are used. In addition, our approach exploits context driven object hierarchy knowledge, where local and object dependent situational information is used for inference on higher object hierarchy levels. This propagation scheme enriches and augments the local deductive reasoning and allows for the injection of asserted activity knowledge for only a subset of the scene objects, while local reasoning can still be performed on higher levels under contextual assumptions. For the motivation and illustration of the applicability and usefulness of our approach, we apply it to human-robot interaction and collaboration scenarios in the industrial domain. A synthetic representation of a shared workspace serves thereby as the basis for the modeling of local object contexts for situational reasoning.