Relationship between UAVs and Ambient objects with threat situational awareness through grid map-based ontology reasoning

Abstract

Autonomous threat situational awareness for unmanned aerial vehicles (UAVs) is required in various fields. Although several approaches have been proposed for autonomous threat situational awareness, most of them involved reasoning of the semantic information of objects. Therefore, in this paper, we propose a method to achieve threat situation awareness in UAVs based on reasoning of the relationship among objects. In this study, there are three main ways to recognize a threat to a UAV. First, information of a recognized objects is expressed using a level of detail-based grid map. Second, the concepts of objects around UAV are defined as an ontology, and the relationship and situations between objects are defined as SWRL. Third, through the ontology reasoning, the simulator visualizes recognizing the relationships of objects and threat situations for the UAV. To validate the proposal, a data generator was constructed to perform reasoning for a virtual threat situation. The generated data ensured that the expected relationships between objects and situations were properly recognized. Four scenarios were analyzed and the relationships and situations that might be encountered in each scenario were set. The accuracy of the overall situational awareness averaged 95%, and the reasoning speed averaged 583 ms.