Report on Updated ASA System Architecture (AWARE Deliverable 2.21)

This document investigates improved options for the architecture of the Artificial Situational Awareness (ASA) system, a core component of the AWARE project, which aims to develop artificial intelligence (AI) support for air traffic controllers, where the AI component has information about task-relevant context and the controller’s attention. Building on prior work in the AISA project, this document explores replacing the original Prolog-based knowledge graph (KG) implementation in the ASA system with either a native RDF-based or a relational implementation, evaluating both in terms of performance, rule expressiveness, and development effort. The document also introduces a cloud-native lakehouse architecture for scalable ingestion and indexing of heterogeneous, high-velocity data sources. This KG Lakehouse allows for on-demand KG generation from the indexed data and is compatible with both a native RDF-based implementation and a relational implementation of KG management for the downstream data processing tasks.

Keywords: Air traffic management, artificial intelligence, knowledge graph management