Offline language tool assists pilots with air combat decisions rapidly
Offline Language-Based Support System for Tactical Decision-Making in Air Combat Operations
Computation and LanguageHuman-Computer Interaction
Summary
Air combat pilots need quick access to tactical information but often rely on slow manual searches in manuals. The authors created MATE, a system that answers pilots' natural language questions using tactical manuals entirely offline, ensuring fast, private support. Tested by Brazilian Air Force pilots, it answered questions with good accuracy and speed using open-source language models. The system could improve pilot training and onboard decision-making by reducing search time.
What this means in practice
- •For military flight trainers: Use MATE to quickly provide tactical answers during pilot training sessions without internet access.
- •For flight simulator developers: Integrate MATE for realistic, instant tactical guidance during simulated air combat exercises.
Authors
Joao P. A. Dantas
Abstract
An offline language-based support tool is presented to assist pilot training and tactical decision-making in air combat operations. The system, named MATE (Mission Assistant for Training and Evaluation), applies retrieval-augmented generation (RAG) to answer natural language questions using content extracted from doctrinal and tactical manuals. Built entirely with open-source components, it performs text preprocessing, embedding generation, semantic search, and language generation locally, ensuring full offline functionality. The tool was evaluated by three Brazilian Air Force fighter pilots, who rated the answers produced with three open-weight language models (Llama 3.1 8B, Qwen 2.5 14B, and Mistral NeMo 12B) to five tactical questions derived from a publicly available air combat manual. Average ratings ranged from 4.00 to 4.40 on a five-point scale, with Mistral NeMo 12B obtaining the highest score, and the pilots rated the response time of the tool as fully adequate compared to manual search. The relevance of the retrieved content was identified as the main point for improvement. Potential applications are discussed in onboard mission systems, flight simulators, and training environments.