An AI-Based Multi-UAV Coordination and Navigation System with Real-Time Visual Recognition and Precision Control

Author : Chang-Hsi Wu, Rui-Ting Jiang, Yu-Heng Kor, Sian-Ci Cheng, Cheng,Yung-Long

Abstract : This study presents an AI-driven multi-UAV system for real-time coordination and navigation. Powered by Jetson Xavier NX and ROS2, the UAVs integrate YOLOv11 for real-time object detection, scene understanding, and mission inference, enabling autonomous target search and vision-based guided landing. RTK-based precision positioning and IMU data support accurate navigation, while a gimbal-mounted camera enables dynamic tracking and visual guidance. The front-end interface, developed using Flutter, supports multi-UAV task control and real-time monitoring. The backend, implemented in Go and integrated with a MySQL database, manages APIs and mission records. Real-time image and command streaming is achieved via WebSocket. By embedding AI-based perception, planning, and decision-making capabilities across multiple UAVs, the system enhances scalability, responsiveness, and control accuracy, providing a robust foundation for autonomous aerial system development

Keywords : AI-driven UAVs, multi-UAV coordination, YOLOv11, ROS2, real-time navigation, autonomous landing, RTK positioning, object detection, Flutter interface, aerial robotics.

Conference Name : International Conference on Computer Science, Technology and Artificial Intelligence (ICCSTAI-25)

Conference Place : Chongqing, China

Conference Date : 28th May 2025