At the Teaching Robotics with ROS 2 workshop during RosCon UK 2025, Prof Marc Hanheide demonstrated how containerised development environments are transforming robotics education. Our L-CAS-developed “Robot in a Box” solution addresses one of the most persistent challenges in robotics pedagogy: getting students from installation frustration to productive learning, very quickly.

To get students to learn from the first minute in our robotics and AI focused programmes, such as our “MSc Robotics and AI” or “BSc (Hons) Computer Science with Artificial Intelligence“, for example, the team at Lincoln has developed a re-usable, containerised approach to deploying robotics softwre for education, which already see adoption by other academic institutions. It’s innovations like this that makes studying robotics and AI at Lincoln somewhat special.

The Challenge

Traditional robotics education faces a familiar hurdle—students often spend more time wrestling with environment setup than learning robotics concepts. The complexity of ROS2 installations, GPU driver conflicts, and platform-specific dependencies can consume days of valuable teaching time. In our experience, traditional setup approaches achieved only ~60% success rates with typical student, with students frequently encountering the dreaded “it works on the lecturer’s machine only” syndrome.

Our Solution: DevContainers for Robotics

L-CAS has developed a comprehensive GPU-accelerated ROS2 development environment delivered as a DevContainer. This containerised solution provides students with a fully configured development environment in under five minutes from initial setup, running seamlessly across Windows, Linux, and macOS hosts.

Some Key Technical Features

• GPU-accelerated 3D visualisation with VirtualGL and NVIDIA runtime support
• Virtual desktop with web-based access via noVNC, eliminating complex local installations
• Cross-platform compatibility ensuring consistent experiences across all student devices
• Zenoh bridging technology enabling seamless connection between containerised environments and real-world robots
• Professional development workflows through integrated Visual Studio Code support

Real-World Impact

Our deployment in our modules has delivered measurable improvements:

• Setup time reduced from 2-4 hours to ~5 minutes
• Success rate increased from ~60% to ~98% of students who wanted to install the system
• Support tickets reduced by an estimated 85%
• Teaching time refocused from IT troubleshooting to robotics content

Live Demonstration

The workshop featured our TidyBot navigation demo, showcasing:

• Real-time SLAM mapping and autonomous navigation
• 3D simulation environments with GPU acceleration
• RViz visualisation and interactive goal setting
• Seamless integration of simulation and real robot connectivity

Students can immediately engage with complex robotics concepts including SLAM, navigation stacks, and sensor fusion—all without the traditional installation barriers.

Open Source and Community Impact

Our entire solution is available as open source, including:

• Complete repository template for immediate adoption available
• Production-ready Docker images hosted at lcas.lincoln.ac.uk
• Comprehensive documentation and deployment guides
• Integration examples for GitHub Codespaces and cloud platforms

Future Directions

We’re expanding support for ARM64 architectures, enhancing cloud platform integration, and developing multi-robot simulation environments. The approach represents a fundamental shift towards consistent, scalable robotics education that can adapt to hybrid and remote learning scenarios. All closely aligned with our research in robotics.

Workshop Resources

The complete demonstration and educational materials from our RosCon25 workshop are available at https://ros2edu.github.io/. We encourage educators to explore this approach and contribute to the growing community of containerised robotics education.

For more information about our work at L-CAS, visit https://lcas.lincoln.ac.uk/.