Embedded reasoning for UAV operations: towards real-time efficiency and trustworthy autonomy

Unmanned Aerial Vehicles are increasingly used in a variety of applications. Operations control currently relies on ground control stations for data analysis and decision-making, incurring in latency issues and requiring expensive computing resources. This paper proposes the integration of Knowledge Representation and Reasoning techniques into UAV autopilot platforms to enable autonomous context management, improved operational efficiency, and real-time risk detection. Computational efficiency and explainability characterize the adopted approach, promoting trustworthiness of autonomous UAV units, as shown by two case study reports.