Robot BMW Develops Software for Humanoids in Landshut

Related Vendors

dataTec AG

Taiwan Machine Tool & Accessory Builders' Association (TMBA)

SIMTOS (Seoul International Machine Tool Show)

BMW is developing software for humanoid robots in component manufacturing at its Landshut (Germany) plant. The robots are intended to perceive their surroundings, assess situations, and derive actions.

BMW is developing software for AI-supported robots in component manufacturing at its Landshut plant. The focus is on open software platforms, training data, simulation and motion planning, as well as training robots. Humanoid robots are intended to independently perform complex tasks in industrial environments. 

To achieve this, the systems should be able to perceive their surroundings, assess situations, and derive appropriate actions. The technical core is an "Intelligence Stack" that combines perception, learning, simulation, and decision-making.

Open platforms are intended to help train robots flexibly and scalably—through a combination of classical, deterministic programming and AI models. These integrate image information, language understanding, and the robot's actions. The goal is a modular robotics ecosystem where different systems work together.

In Landshut, developers are working on determining which tasks are suitable for humanoid robots. This mainly involves activities that require flexibility, fine motor skills, and an understanding of the working environment. These include recognizing and categorizing environments, determining position and condition, planning efficient movements, and precisely grasping and moving components. Landshut is the competence center for in-house component manufacturing at BMW.

Robotics in Automotive Manufacturing

Humanoid Robot Will Sort Components at BMW in the Future

Robots Learn by Demonstration

At the plant, the project team collects data from real production environments and test setups to integrate it into robot simulations. The system virtually maps motion sequences and motion planning; developers can pre-test processes and experiment with new approaches outside ongoing production.

In parallel, humanoid robots are acquiring new skills through demonstration-based learning. This means the project team demonstrates motion sequences and work processes. The developers capture the resulting data using cameras, motion-capture suits, and data gloves—and convert it into generalizable behavioral models, known as policies. These policies can be applied to different tasks and applications.

Development with Athenyx Robotics

One of the partners in the project is Athenyx Robotics, a startup established specifically for this technology area. It originates from the Machine Tool Laboratory of RWTH Aachen (Germany) and is located in close proximity to the BMW Landshut plant.

"For us, it is crucial to develop the technology very close to real requirements. We are not only examining what is technically possible but also where humanoid robotics can actually provide a tangible benefit in component manufacturing," explains Christoph Jagoda, Project Manager Physical AI at the BMW plant in Landshut. "This way, a clear picture gradually emerges of which applications are suitable for industrial use."

The platform-based approach aims to enable the use of learned skills across robotic systems and application cases. Additionally, the Landshut plant collaborates with the University of Landshut, for example, in integrating data, simulation, and developing and testing specific applications. The project complements ongoing pilot projects on the use of humanoid robotics at the plants in Leipzig (Germany) and Spartanburg, USA.