Aerial Environment and Rules Ontology(AERO)

The increasing complexity of low-altitude airspace necessitates intelligent agents (such as UAVs) that possess a deep, computable understanding of their operational environment. However, the required knowledge is often fragmented across unstructured physical, informational, and legal/regulatory domains, posing a significant challenge to consistent semantic interpretation and safe decision-making. To address this, we introduce AERO (Aerial Environment and Rules Ontology), a comprehensive semantic framework, and a novel workflow for its construction. Our primary contribution is a systematic methodology that uses a multi-agent LLM system to semi-automatically build the domain-specific ontology and populate a large-scale Aerial Environment Knowledge Graph (AEKG) from heterogeneous sources. The resulting AERO ontology provides a standardized vocabulary that unifies airspace rules, dynamic entities (e.g., weather, aircraft), and geospatial features, while the AEKG serves as a precise, context-rich knowledge base for downstream applications. To validate our approach, we demonstrate the AEKG's utility by integrating it into a Retrieval-Augmented Generation (RAG) system, showing a significant improvement in answering complex spatiotemporal queries and a reduction in factual hallucinations compared to traditional document-based RAG. Our work provides both a foundational knowledge base and a scalable methodology for enhancing the environmental cognition and decision-making capabilities of low-altitude intelligent agents.