Tutors: Carlos Medina and Simon Janzon
Join Our Bachelor Project in Fleet Robotics, AI, and 3D Scene Understanding
Are you ready to build not just one autonomous robot, but an entire team of robots that can understand tasks, coordinate their actions, and operate intelligently in shared environments?
FleetGraph is a hands-on bachelor project combining multi-robot systems, robotics, computer vision, artificial intelligence, web technologies, and 3D scene graphs. You will develop a complete fleet-management platform that receives tasks in plain language, assigns them to suitable robots, monitors their execution in real time, and enables robots to share semantic information about the environment.
What You’ll Build
You will work with a fleet of simulated mobile robots. The system should be able to:
-
Receive tasks such as: “Deliver a package to laboratory 2.”
-
Assign tasks according to robot position, battery, workload, and capabilities.
-
Visualize robots and tasks in a real-time dashboard.
-
Use 3D scene graphs to represent places, obstacles, objects, and people.
-
Support socially aware navigation around people and groups.
-
Share observations between robots to update a common scene representation.
The Challenge: Teaching Robots to Understand Affordances
A robot should not simply choose the shortest path or the closest task. It must consider the current state of the fleet and the social context of the environment.
Your mission is to develop a system that can:
-
Coordinate multiple robots.
-
Recover from robot failures and task reassignment.
-
Use scene graphs for semantic navigation.
-
Share information between robots.
-
Navigate safely in human-populated environments.
-
Evaluate whether semantic and social information improves fleet performance.
The project is suitable for bachelor students interested in robotics, AI, computer vision, software engineering, or web technologies. Previous experience with ROS, robotics, or computer vision is helpful but not required.
The course may be held in English and runs for one semester. The number of places is limited to 12 places (initially), and participation in the kick-off meeting is mandatory.
The information about the kick-off meeting will be published in the upcoming weeks.
For questions and registration, please contact carlos.medina-sanchez@uni-due.de.