Investigating Large Language Models as Collaborative Teammates for Human-AI Aerial Teaming Scenarios

This project investigates the use of large language models (LLMs) as autonomous teammates to perform aircraft control, strategic reasoning, and team communication in complex aviation missions. We propose to design and evaluate an LLM-based “wingman” agent for a simulated collaborative aerial firefighting scenario, in which a human and AI fly their own aircraft to detect and extinguish wildfires. Unlike prior work that employs narrow AI systems to fulfill isolated teamwork functions, this research explores whether a single LLM can integrate real-time reasoning, mission planning, and natural language communication to enable effective human–AI teaming. This work represents the first systematic exploration of LLMs for both autonomous flight control and human collaboration, advancing the integration of language-based intelligence into mission-critical domains. Our research questions focus on (1) the feasibility of LLMs serving as general-purpose teammates in fast-paced, safety-critical aviation missions, and (2) the design choices (such as system prompting, context management, and retrieval-augmented grounding) that most influence performance, adherence to mission rules, and human trust.

Faculty Supervisor:

Ali Ayub

Student:

Partner:

Georgia Institute of Technology

Discipline:

Computer science

Sector:

Aerospace; Artificial Intelligence

University:

Concordia University

Program: