3. Natural language solution for historical meteorological information (‘ChatMET’)

Accessing meteorological data is fundamental to post-operational analysis, safety investigations, and airspace design within Air Traffic Management (ATM). However, current processes are often manual, inefficient, and require significant human interaction between ATM and meteorological services. The ChatMET project was initiated to investigate the feasibility of leveraging Large Language Models (LLMs) to create an intuitive, natural language interface for accessing these complex datasets. The core of the research focused on addressing the inherent reliability challenges of LLMs, such as model hallucinations. This investigation led to the identification and validation of a “tool-calling” agentic architecture where the LLM acts not as a source of knowledge, but as a secure translator that converts user queries into API calls to an authoritative source of truth. The project’s primary outcome is a Technology Readiness Level (TRL) 3 proof-of-concept that serves as a foundational blueprint for the future development of trustworthy AI-assisted tools in ATM.