Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Agentic AI Meets Edge Computing in Autonomous UAV Swarms

Thuan Minh Nguyen, Vu Tuan Truong, Long Bao Le

TL;DR

The paper tackles the challenge of scalable, resilient UAV swarm autonomy in dynamic environments by integrating agentic AI with edge computing. It articulates three deployment architectures (standalone, edge-enabled, and edge/cloud-enabled) and demonstrates the edge-enabled design through a wildfire SAR use case with on-board TinyLLaMA planning and edge-LLM validation. Through LangGraph-based multi-agent frameworks and an edge-ground-station workflow, the study shows improved coverage and reduced mission times relative to a state-of-the-art baseline, while addressing hallucinations via prompt validation. The paper also discusses open challenges—efficient onboard LLMs, hallucination control, robust collaboration, scalable edge infrastructure, and UAV hardware constraints—outlining a path toward practical, mission-critical swarm operations.

Abstract

The integration of agentic AI, powered by large language models (LLMs) with autonomous reasoning, planning, and execution, into unmanned aerial vehicle (UAV) swarms opens new operational possibilities and brings the vision of the Internet of Drones closer to reality. However, infrastructure constraints, dynamic environments, and the computational demands of multi-agent coordination limit real-world deployment in high-risk scenarios such as wildfires and disaster response. This paper investigates the integration of LLM-based agentic AI and edge computing to realize scalable and resilient autonomy in UAV swarms. We first discuss three architectures for supporting UAV swarms - standalone, edge-enabled, and edge-cloud hybrid deployment - each optimized for varying autonomy and connectivity levels. Then, a use case for wildfire search and rescue (SAR) is designed to demonstrate the efficiency of the edge-enabled architecture, enabling high SAR coverage, reduced mission completion times, and a higher level of autonomy compared to traditional approaches. Finally, we highlight open challenges in integrating LLMs and edge computing for mission-critical UAV-swarm applications.

Agentic AI Meets Edge Computing in Autonomous UAV Swarms

TL;DR

The paper tackles the challenge of scalable, resilient UAV swarm autonomy in dynamic environments by integrating agentic AI with edge computing. It articulates three deployment architectures (standalone, edge-enabled, and edge/cloud-enabled) and demonstrates the edge-enabled design through a wildfire SAR use case with on-board TinyLLaMA planning and edge-LLM validation. Through LangGraph-based multi-agent frameworks and an edge-ground-station workflow, the study shows improved coverage and reduced mission times relative to a state-of-the-art baseline, while addressing hallucinations via prompt validation. The paper also discusses open challenges—efficient onboard LLMs, hallucination control, robust collaboration, scalable edge infrastructure, and UAV hardware constraints—outlining a path toward practical, mission-critical swarm operations.

Abstract

The integration of agentic AI, powered by large language models (LLMs) with autonomous reasoning, planning, and execution, into unmanned aerial vehicle (UAV) swarms opens new operational possibilities and brings the vision of the Internet of Drones closer to reality. However, infrastructure constraints, dynamic environments, and the computational demands of multi-agent coordination limit real-world deployment in high-risk scenarios such as wildfires and disaster response. This paper investigates the integration of LLM-based agentic AI and edge computing to realize scalable and resilient autonomy in UAV swarms. We first discuss three architectures for supporting UAV swarms - standalone, edge-enabled, and edge-cloud hybrid deployment - each optimized for varying autonomy and connectivity levels. Then, a use case for wildfire search and rescue (SAR) is designed to demonstrate the efficiency of the edge-enabled architecture, enabling high SAR coverage, reduced mission completion times, and a higher level of autonomy compared to traditional approaches. Finally, we highlight open challenges in integrating LLMs and edge computing for mission-critical UAV-swarm applications.
Paper Structure (25 sections, 2 equations, 5 figures, 1 algorithm)

This paper contains 25 sections, 2 equations, 5 figures, 1 algorithm.

Table of Contents

  1. Introduction
  2. Agentic AI Systems
  3. Background on Agentic AI
  4. Multi-agentic AI Frameworks
  5. UAV Swarm Design with LangGraph
  6. Multi-Agentic AI for UAV Swarms
  7. Standalone UAV Swarm
  8. Edge-enabled UAV Swarm
  9. Edge/Cloud-enabled UAV Swarm
  10. Applications
  11. Use Case: Wildfire Search-and-Rescue (SAR)
  12. Mission Objectives
  13. System Architecture and Operational Workflow
  14. Performance Metrics
  15. State-of-the-Art Baseline
  16. ...and 10 more sections

Figures (5)

  • Figure 1: Three multi-agentic AI deployment architectures for UAV swarms: (a) Standalone UAV swarm; (b) Edge-enabled UAV swarm; (c) Edge/cloud-enabled UAV swarm; (d) Applications; (e) UAV agent
  • Figure 2: Proposed multi-agentic AI-driven UAV swarm system for wildfire SAR.
  • Figure 3: Visualization of survey-point assignments for 10 UAVs across three different phases, shown alongside the expanding wildfire region.
  • Figure 4: Coverage rate of proposed design and the baseline [13]
  • Figure 5: Mission completion time of proposed design and other baselines