Recently, the arculee brain was migrated to ROS 2 (Robot Operating System 2) to ensure safer, reliable, and efficient operation. In this blog post, Software Engineer Thomas Fuhrmann shares the hows and whys behind this move.

At arculus, we implemented one of the most significant upgrades in our robotics journey: migrating the arculee brain to ROS 2 (Robot Operating System 2). ROS 2 is an open-source robotics framework designed for flexibility, scalability, and safety, which are the three core principles that guide everything we do.

Thomas Fuhrmann, our Software Engineer, elaborates on why ROS 2 is such a significant shift:

“It’s not just a technical update. It’s a complete re-architecture of how our robots think, act, and adapt.”

Splitting the Robot Brain

The current arculee brain is powerful but also incredibly intricate; everything from navigation to safety to customer-specific hardware lives in one system. With ROS 2, we simplify it by splitting it into two major layers:

1. Autonomy Layer: Handles navigation, path planning, obstacle avoidance, and executing orders. This is where the robot's intelligence resides.
2. Platform Layer: Handles everything specific to a particular robot, including its sensors, drive system, and any unique configurations.

Thomas explains the key benefit of this division of layers,

“For instance, if the next generation of our robots comes with a different mechanical design, we would just need to exchange the platform side. So, develop it anew and keep the whole autonomy stack, making our development much faster.”

ROS 2 = Safer, Smarter and more Flexible Robots

ROS 2 has different benefits that can help address some of the past challenges and enhance the arculee Brain. Thomas highlights the following key aspects:

Decoupling for Safety

At arculus, safety is non-negotiable. Every arculee has special onboard hardware dedicated to preventing accidents and protecting humans. But that safety logic was tightly woven into the rest of the brain’s code.

Migrating to ROS 2 means decoupling safety from autonomy logic, keeping it as an independent module. This separation makes the system more reliable and ensures that safety-critical features stay isolated from other functions.

Scalable Robotics with Modular Backpacks

Customer needs sometimes vary across different transportation use cases. This is where backpacks come in, offering flexibility to adapt accordingly. These are physical modules added on top of an arculee to handle specific payloads, like pallets, turntables, or more advanced custom tools.

Previously, the integration of new, more complex backpacks may not always be as smooth as desired. However, ROS 2’s modular architecture makes it easier to adapt, ensuring we can quickly support new customer requirements without having to reinvent the wheel each time. This flexibility is essential as we scale to more customers with increasingly specialised demands.

Smarter Fault Management

Robots, like any complex system, don’t always behave perfectly. So, in case of a problem, the go-to method was to shut everything down and restart. This approach works, but it’s not ideal.

With ROS 2, we can take advantage of its lifecycle management. That means instead of rebooting the entire brain, we can restart only the faulty part of the system, thus minimising downtime and keeping operations running smoothly.

Discussing modularity, Thomas also shares that the team is moving towards Docker-based deployments. He believes that it would allow:

• Faster testing by isolating and validating individual modules
• Easier scaling to different environments and customer needs
• More robust and maintainable software overall

Why It All Matters

Migrating to ROS 2 isn’t just about rewriting code; it’s about rethinking how our robots evolve. By splitting autonomy from the platform, decoupling safety, improving backpack integration, and enabling smarter fault management, arculus is building a future-ready foundation. The arculee brain becomes more flexible, safer, and scalable than ever.

Learn more in the video below!