AgriRobot: Autonomous Navigation Using ArUco Vision and ROS2
Student: Mohamed MTIBAA
PFE 2026
ENET’COM / Industrial Computer Engineering
This project develops a low-cost autonomous navigation system for the AgriRobot agricultural robot using ArUco vision and ROS2. It enables reliable movement inside greenhouses for smart agriculture applications.
Project Highlights:
ArUco-Based Navigation
Use of visual ArUco markers with a Raspberry Pi Camera Module 3 for robot localization, pose estimation, and guidance inside the greenhouse.
Use of visual ArUco markers with a Raspberry Pi Camera Module 3 for robot localization, pose estimation, and guidance inside the greenhouse.
ROS2 Modular Architecture
Development of a ROS2 Humble architecture composed of nodes for video acquisition, marker detection, pose estimation, mission management, and motion control.
Development of a ROS2 Humble architecture composed of nodes for video acquisition, marker detection, pose estimation, mission management, and motion control.
Embedded Low-Cost Platform
Deployment on a Raspberry Pi 4 running Debian GNU/Linux ARM64, with ESP32-based motor control and open-source software tools.
Deployment on a Raspberry Pi 4 running Debian GNU/Linux ARM64, with ESP32-based motor control and open-source software tools.
Real-Time Performance
Optimization of the vision pipeline using multiprocessing, reaching 30 FPS and a marker detection rate above 99% during tests.
Optimization of the vision pipeline using multiprocessing, reaching 30 FPS and a marker detection rate above 99% during tests.
Autonomous Mission Execution
Management of waypoint sequences, autonomous movement through crop rows, and automatic return to the charging station.
Management of waypoint sequences, autonomous movement through crop rows, and automatic return to the charging station.
Greenhouse Validation
Validation of the navigation system in representative agricultural conditions, demonstrating a reliable and economical solution for protected environments.
Validation of the navigation system in representative agricultural conditions, demonstrating a reliable and economical solution for protected environments.
Prototype
The prototype demonstrates the autonomous navigation capabilities
of AgriRobot in a greenhouse environment. It illustrates how the
robot moves through predefined paths while maintaining accurate
navigation and avoiding navigation errors.
This prototype validates the feasibility of the proposed solution and highlights its potential for real agricultural applications.
This prototype validates the feasibility of the proposed solution and highlights its potential for real agricultural applications.
