Projecte llegit
Títol: Distributed architecture for real-time UAV-UGV coordination
Estudiants que han llegit aquest projecte:
PÉREZ PÉREZ, JOAN (data lectura: 13-07-2026)- Cerca aquest projecte a Bibliotècnica
PÉREZ PÉREZ, JOAN (data lectura: 13-07-2026)- Cerca aquest projecte a Bibliotècnica
PÉREZ PÉREZ, JOAN (data lectura: 13-07-2026)
PÉREZ PÉREZ, JOAN (data lectura: 13-07-2026)Director/a: LÓPEZ RUBIO, JUAN
Departament: DAC
Títol: Distributed architecture for real-time UAV-UGV coordination
Data inici oferta: 28-01-2026 Data finalització oferta: 28-09-2026
Estudis d'assignació del projecte:
DG ENG AERO/TELEMÀT
| Tipus: Individual | |
| Lloc de realització: EETAC | |
| Paraules clau: | |
| Drone, transmissió, temps real, mqtt | |
| Descripció del contingut i pla d'activitats: | |
| L'objectiu d'aquest projecte es desenvolupar sistema per enviar dades telemetria e imatge d'un drone a un sistema de terra. Es fara servir Mavlink, MQTT i ROS 2 per fer simulacions realistes de l'entorn. | |
| Overview (resum en anglès): | |
| This project presents the design, implementation and validation of a distributed architecture for real-time coordination between an unmanned aerial vehicle (UAV), a ground robot (UGV) and a browser-based supervision service within the 6G OpenLab infrastructure. The main objective is to connect aerial sensing, terrestrial actuation and remote supervision through modular communication flows that can operate in both simulated and physical environments. To achieve this, the SafeFlight web application was developed to command the UAV and display its telemetry and live video. In parallel, an MQTT-to-ROS translation bridge was implemented to transform MAVLink telemetry into JSON messages distributed through MQTT and translated into ROS for the Robotnik RB-SUMMIT ground robot.
The methodology followed an evolutionary development process for each individual component of the project. Individual parts of the distributed architecture were first implemented and tested independently, including ArduPilot SITL simulation, MAVLink communication, MQTT publication and subscription, robot control, multimedia transport and the SafeFlight interface. These components were progressively integrated and validated through simulation, local tests, deployment and experiments at the DroneLab. The architecture separates telemetry, command and multimedia traffic according to their requirements. Direct MAVLink communication is used for UAV supervision and control, MQTT is used for lightweight machine-to-machine data distribution, ROS manages robot navigation, and agnostic-to-camera video is exposed to browsers through MediaMTX. For development and deployment, tools such as Docker, Dev Containers, CI/CD workflows and Kubernetes were used. The results demonstrate that the same upper-layer communication interfaces can be maintained when moving from simulation to a CubePilot-based UAV. SafeFlight worked successfully, allowing flights from anywhere within the infrastructure. The middleware correctly converted MAVLink data into MQTT JSON payloads and forwarded the selected geographical goals to the robot control layer. The Robotnik platform received the objectives and generated ROS velocity commands through a closed-loop navigation controller. The complete multimedia chain was also validated, from H.264 stream generation to WebRTC delivery in SafeFlight. Overall, the project provides a functional and modular foundation for future UAV-UGV cooperation scenarios, particularly in precision agriculture, while identifying future work in autonomous detection and multi-UAV operation. |
|