The European Robotics Forum traditionally offers a unique combination of academic insights and practical robot applications. This year, partly due to the prominent presence of the automotive industry in host city Stuttgart, there will be even more focus on the practical deployment of robots. Media partner Rocking Robots spoke with Wesley Roozing from the University of Twente, who is responsible for two of the workshops.
“Dirty, wet, hot, cold, vibrations, collisions, high radiation—these are often the environments where robots are deployed,” says Wesley Roozing, Assistant Professor of Robotics and Mechatronics at the University of Twente. He chairs the Topic Group on Mechatronics at euRobotics. One of the workshops he is involved in at ERF focuses specifically on robustness, resilience, and reliability in mechatronics, or as he puts it, “how to deploy robots in the real world.”
Roozing explains: “Take inspection robots, for example. They often need to function underwater or withstand sand, dust, and even radiation, such as at CERN. These robots must be able to operate robustly, without suffering from component failures, and if failures do occur, they must be resilient enough to handle minor malfunctions.”
Academic Approach
This represents a different approach than in academia, where research is often prioritized over building robots that can endure for long periods. “That is an important additional step: creating a robot that remains reliable in the long term and is resistant to external influences. One of the main challenges is the transition from a controlled environment to an unpredictable, often hostile setting.”
“In a laboratory, conditions are stable—temperature, humidity, and even the surface a robot moves on are predictable. But once deployed outdoors or in an industrial setting, other factors come into play. This means both the hardware and software must not only function under optimal conditions but also withstand disruptions and unexpected situations.”
Nuclear Power Plants
Reliability is another major challenge. “In a laboratory, components can be easily replaced, or a robot can be reset if something goes wrong. In practice, that is often not possible. Consider robots performing maintenance on offshore drilling platforms or in nuclear power plants. They need to function over long periods without constant human intervention. This requires robust designs, redundancy in critical systems, and self-diagnosis mechanisms so the robot can detect problems and, if necessary, recover or adopt an alternative approach.”
Another important aspect is interaction with both the environment and humans. “In a laboratory, a robot’s sensors and algorithms can be tested under ideal conditions, but in the real world, there are unexpected objects, dynamic environments, and people who behave unpredictably. This requires more advanced perception techniques and decision-making algorithms that can respond in real-time to changing situations without the robot becoming stuck or making incorrect decisions.”
Finally, there is the challenge of scalability and cost. “Many techniques developed in research are not immediately applicable on a large scale. Materials, components, and production methods must be economically viable for a robot to be deployed in industrial or commercial applications.”
Humanoid Robots
These issues also relate to two other workshops Roozing is involved in, which focus on humanoid robots. “A humanoid robot working on a production line needs to function reliably for more than just a few days or weeks and must integrate into an environment that is very different from a lab. Human coworkers will likely be present, meaning the requirements are entirely different.”
“Moreover, a humanoid robot is one of the most complex types of robots to build. It has a large number of joints, which brings significant costs and complexity. Each joint requires actuators, sensors, motors, gearing, and controllers, all of which can fail. This presents numerous challenges.”
The Robotic Hand
One of the most complex parts of a robot is its hand. “I don’t think we are even close to replicating the human hand, which is a biological marvel. Of course, for specific use cases, we have effective grippers, but general-purpose hands like ours, with five fingers and associated flexibility and tactile sensitivity, remain unmatched.”
“There are a few highly complex and therefore expensive robotic hands that come somewhat close in terms of range of sensing and the number of joints. But even then, their complexity makes them extremely difficult to use. We still don’t have a good way to control them so that a robot can grasp objects as robustly as a human without visual feedback.”
Computer Vision
Perceiving the environment is another crucial aspect of robotics. “Computer vision is not my area of expertise, but I believe we have made significant progress. Robots are quite good at recognizing objects, but understanding how those objects relate to one another is much more difficult. For example, if a bottle cap is lying on a table next to a bottle, making the connection that the cap should be screwed onto the bottle is not always straightforward for a robot. That is the cognitive aspect of vision.”
“Furthermore, robot vision systems lack a fundamental understanding of physics. If a robot sees an object that appears unstable, humans can immediately recognize this and predict what will happen if force is applied. A robot does not have this awareness.”
Practical Applications
Various sectors are actively experimenting with the deployment of such robots. At ERF 2025, companies including BMW and Mercedes will share their experiences and the challenges they face. “Ultimately, the use of robots is driven by cost and productivity. If their productivity approaches that of human workers, then there is a business case for humanoid robots. At the moment, they are still slower than humans, and their acquisition and operational costs are high.”
He concludes: “With topics like these, ERF is a unique event. Compared to purely scientific conferences, it places much more emphasis on strategy, applications, and business. This gives attendees a clearer view of the broader and more pressing challenges in robotics.”
“This is also reflected in the participants: alongside academics and researchers, there are many professionals from the business sector, including both users and developers of robotic hardware. Additionally, representatives from the European Commission attend, as they ultimately decide the funding for new robotics research. This makes ERF a leading conference for shaping long-term strategies.”
Registration for ERF 2025 is still open
