This paper presents the LeRAAT framework, which integrates a large-scale language model (LLM) with the __T13489_____-Plane flight simulator to provide real-time situational awareness-based pilot assistance. LeRAAT uses real-time flight data, weather conditions, and aircraft documentation to generate recommendations that are consistent with aviation best practices and tailored to specific situations. It uses a search-augmented generation (RAG) pipeline that extracts and synthesizes information from aviation regulatory sources such as aircraft type-specific manuals (including performance specifications and emergency procedures), FAA instructions, and standard operating procedures. The framework is demonstrated in both virtual reality and conventional screen simulations, supporting a wide range of research applications such as pilot training, human factors research, and operational decision support.
