Aero-LLM: A Distributed Framework for Secure UAV Communication and Intelligent Decision-Making
Balakrishnan Dharmalingam, Rajdeep Mukherjee, Brett Piggott, Guohuan Feng, Anyi Liu
TL;DR
Aero-LLM addresses the need for secure, reliable UAV communication and intelligent decision-making by deploying a distributed team of specialized LLMs across onboard, edge, and cloud levels. The framework uses task-specific fine-tuning through supervised fine-tuning and reinforcement learning from human feedback, with TimesNet for anomaly detection and Time-LLM for forecasting, validated on network, sensor, and cyber-security tasks. The results show strong task-specific performance with robust security resilience and manageable memory footprints, achieved via a hierarchical, resource-aware deployment. This approach advances secure, scalable, and autonomous UAV operations, enabling resilient operation in diverse environments and paving the way for multi-UAV collaboration and real-time data-driven decision-making.
Abstract
Increased utilization of unmanned aerial vehicles (UAVs) in critical operations necessitates secure and reliable communication with Ground Control Stations (GCS). This paper introduces Aero-LLM, a framework integrating multiple Large Language Models (LLMs) to enhance UAV mission security and operational efficiency. Unlike conventional singular LLMs, Aero-LLM leverages multiple specialized LLMs for various tasks, such as inferencing, anomaly detection, and forecasting, deployed across onboard systems, edge, and cloud servers. This dynamic, distributed architecture reduces performance bottleneck and increases security capabilities. Aero-LLM's evaluation demonstrates outstanding task-specific metrics and robust defense against cyber threats, significantly enhancing UAV decision-making and operational capabilities and security resilience against cyber attacks, setting a new standard for secure, intelligent UAV operations.