All-rounder brings smart factory to next level

“O³dyn”: simulation-based artificial intelligence

October 05, 2023

With its reduced-infrastructure and automated logistics, the smart factory is no longer just a vision. It has long been a reality in the automotive industry and among its suppliers. Now, researchers at the Fraunhofer Institute for Material Flow and Logistics IML are demonstrating how simulation-based artificial intelligence can take the smart factory to the next level. With the highly dynamic autonomous transport robot “O³dyn,” which debuted at LogiMAT in Stuttgart, the future of material flow has arrived.

As an autonomous transport robot, O³dyn is one of the first inhabitants of the “robotic continuum,” which merges simulation and machine learning with reality. With the highly dynamic autonomous transport system, the Dortmund-based researchers intend to significantly transform external and internal pallet handling. “O³dyn” was developed as part of the major Silicon Economy research project. With this project, which received more than 25 million euros in funding over three years from the German Federal Ministry for Digital and Transport (BMDV), Fraunhofer IML aims to bring about a breakthrough in Germany and the rest of Europe with a decentralized, federal and open platform economy.

Greater security and efficiency in pallet handling

Privacy warning

With the click on the play button an external video from is loaded and started. Your data is possible transferred and stored to third party. Do not start the video if you disagree. Find more about the youtube privacy statement under the following link:

Unlike most driverless transport systems, which provide either high performance, dynamics or flexibility and are designed for either indoor or outdoor use, “O³dyn” combines all three properties in one. Not only is it highly dynamic and autonomous; it is also suitable for hybrid use. At speeds of up to 36 km/h, it can transport heavy loads in pallet format omnidirectionally. In doing so, it leaves the protected and defined environment of warehouses to operate dynamically on company premises. That is why the researchers call it “O³dyn” (pronounced “Odyn”), which stands for omnidirectional, outdoor and open source.

Odyn really comes into its own when materials need to be moved from one building to another over longer distances on large factory sites. It transports large pallets with ease and moves seamlessly from indoor to outdoor areas. Where forklifts or tugger trains are currently used for tasks of this kind, Odyn could increase resilience and flexibility in the future. Compared to conventional transportation vehicles, Odyn could, according to the researchers, reduce the risk of accidents and significantly improve efficiency. A tugger train, for example, always needs to be built first: Parts need to be buffered and put together, then spread along a production line in a milk run. A large amount of work is involved in coordinating this. Instead, individual, flexible and autonomously operating vehicles could travel directly to where they are needed. Odyn’s payload is currently 350 kg. However, the vehicle can easily be set up for heavier payloads. “For industrial use, we could make the frame from steel instead of aluminum. The drive system and chassis are already designed for a weight of 1.3 metric tons,” says Guido Follert, Head of Department Machinery and Systems at Fraunhofer IML.

Omnidirectional chassis for precise load handling

To allow Odyn to travel in indoor and outdoor areas easily, the researchers have combined the omnidirectional chassis with Mecanum wheels and air suspension. This means that the vehicle can not only operate in the smallest of spaces, thanks to traversing and sideways movement, but can also position the load with precision. The chassis adapts to uneven surfaces in outdoor areas. This is thanks to the special wheels and air suspension, which guarantee secure movement on unlevel ground regardless of the load being carried. This protects both the load and the vehicle. The air suspension also allows the vehicle to be loaded by lowering it to pick up a pallet. The load is secured by catches that move in between the pallet blocks and hold them in place. This prevents the pallet from falling off while the vehicle is moving. The braking system consists of an electric service brake and an emergency brake, which operates using friction brake plates under the frame. If the emergency brake is triggered, the air valves on the chassis open. The vehicle lowers itself onto the brake plates and comes to a standstill immediately

© Fraunhofer IML

Smooth navigation between indoor and outdoor environments

Another obstacle that the Fraunhofer research team had to overcome when developing Odyn was to ensure that it could navigate smoothly between indoor and outdoor areas. They solved the problem with localization algorithms based on the environment and radio signals. The vehicle uses lidar scanners, 3D camera systems and a differential GPS/GNSS to establish its location. Another challenge that still remains to be solved is safe autonomous operation. “Like the automotive industry, we too need to find solutions to ensure safe autonomous driving in public spaces, with all the unpredictable factors involved, such as the sudden appearance of obstacles in the vehicle’s path,” explains Guido Follert. “There is still some work for us to do in this area.”

Outdoor AGV
© Fraunhofer IML – Michael Neuhaus
O³dyn, Transportroboter, autonomes Fahrzeug
© Fraunhofer IML – Michael Neuhaus
© Fraunhofer IML - Michael Neuhaus

Into the “robotic continuum” with simulationbased artificial intellig

Fraunhofer IML has developed this nimble transport robot with the help of a new branch of research: simulation-based artificial intelligence. Thanks to modern graphics cards, the Omniverse robotics platform from NVIDIA and the Isaac simulation software, also from NVIDIA, highly complex processes can be simulated in real time. With motion capturing, the scientists can align the behavior in the simulation with that of the real vehicles, allowing them to optimize the simulation model. “The smaller the difference between the model and reality, the more the robot becomes the cyberphysical twin of the simulation,” says Dr. Sören Kerner, Head of Department AI and Autonomous Systems at Fraunhofer IML. “This approach enables us to reduce development times massively.” In this way, prototypes can be tested in digital reality before they are even built. This allows hardware and software development to be decoupled from each other. A digital continuum of development (“robotic continuum”) is emerging. Odyn’s concept and design are being made available as open source applications at the Open Logistics Foundation . Fraunhofer IML has already made contact with the first project partners in the industry.