How Schunk optimizes systems with the metaverse

Mr. Gessmann, isn't "Metaverse" a typical buzzword?

Timo Gessmann: No. At Schunk, we work with the Metaverse every day and will do so even more intensively in the future.

How does Schunk work with it?

Gessmann: We use the power of software and simulation. This makes it easier for our customers to get started with automation. We offer them a mature and highly productive solution very quickly, optimized for their manufacturing tasks.

Why is the Metaverse so important for your company?

Gessmann: The Metaverse provides an answer to the question: How do we automate flexible production? And this question is central for us as a pioneer in clamping technology, gripping technology, and automation technology. We observe in all industries that industrial production is changing, and so are the requirements of customers for our products.

In addition to high-volume and low-mix, the market for high-mix and low-volume is growing. A classic but increasingly rare case is that a plant always produces the same component in large quantities. There, robots and grippers repeat the same movement over and over. In such a case, it is relatively easy to plan in advance using CAD data what the plant will look like and how the components will interact. However, the requirements are often different now.

What are these changed requirements?

Gessmann: Variety of variants. A plant must be able to produce product A today, product B tomorrow, and C the day after tomorrow. Many different manufacturing aspects and variables come together. For example: Can the same gripper gently grasp small parts, but powerfully grasp large steel parts? How fast is the travel speed, how quickly does the system respond? If you build and test all this in the physical reality, it takes a lot of time and energy. In the metaverse, it goes much faster.

What is easier here than in reality?

Gessmann: There, we can quickly and with little effort design complex systems and then test and simulate them in numerous variants. Additionally, we optimize the systems for productivity in the metaverse before they are even built. This means that our customers not only receive their production systems much faster but also start them up highly productively right away.

How does Schunk enter the metaverse?

Gessmann: In five stages. Each one brings us a step closer to the perfect digital twin. We have already completed stage one: CAD models for digital engineering are available for all our 13,000 physical components. However, this is not enough for the simulation of flexible production. That's why we have already brought our most important mechatronic products to stage two. Here, electrical properties such as power consumption or voltage are stored in the digital twin. But also interfaces such as compressed air consumption or sensors. It becomes particularly exciting from stage three onwards.

What happens at this stage?

Gessmann: The communication interfaces are also stored here. But above all, it is about products that use artificial intelligence. Let me give you an example: The 2D Grasping Kit from Schunk is an application solution for gripping and sorting various unsorted components.

It is an easy entry into automation because the camera system independently recognizes the components and their position and calculates the best grip points. Customers need no prior knowledge of programming or image processing. This only works because we previously trained the AI of our 2D Grasping Kit based on the specific task in the metaverse.

What does the AI train on? 

Gessmann: Theoretically everything - that's the beauty of it! In the metaverse, we can train all variants without much effort. The AI learns, for example, to recognize different components: Is it a metal screw or a glass tube? We train different positions, lighting situations, degrees of contamination, possible damage to the component, and so on - all at the push of a button.

We describe and simulate the physical process in the metaverse to subsequently validate the best result in the real world. For example, in one of our 15 CoLabs worldwide - these are Schunk's robot application centers.

What goals has Schunk set in the metaverse?

Gessmann: We want to perfectly simulate the complete physics of our real automation solutions. These are stages four and five. At the Hannover Messe 2024, we presented a digital twin of an automation cell at stage five; based on the Nvidia Omniverse platform.

The use case of the simulation is "gripping and feeding." The digital twin maps physical properties such as speed, frictional force, adhesive force, and much more. The simulation even considers that lubricating grease in the system varies in ease of movement depending on age and temperature, thus affecting the processes.

Is such a deep simulation already operational?

Gessmann: Not yet for a real industrial roll-out. There are still too many weaknesses in micrometer accuracy or long-term simulation - this is referred to as the so-called Sim2Real gap. But the technologies around the industrial metaverse are developing very quickly.

At Schunk, we want to help shape this development. And I also want to encourage other companies to venture into this topic. Go in boldly! Because do you know what the most important thing is?

What is it?

Gessmann: Cooperation. At Schunk, we rely on collaboration, build networks, and actively participate. We have learned that you can progress faster together with others - this is especially true for technological future topics. For example, we work with many companies in the field of AI and are involved in the IPAI, the future European competence center for applied AI in Heilbronn.

We achieve progress here because we trust each other, exchange ideas, and bring AI into application together. This way, we all move forward faster. It will be the same with the metaverse.

Speaking of progress: Where do you see future opportunities to use metaverse technologies?

Gessmann: The fully simulated smart factory will come. And beyond the narrow production simulation, there is of course still much in the business environment that can be digitally played out and improved. I'm thinking of core processes like development and sales. But it doesn't stop there. What about the not-so-easily tangible knowledge in the minds of employees? I can imagine ways to digitize process knowledge acquired over many years by experts using AI. In Europe, many people are currently retiring. This could be a way to retain their knowledge and experience in the company.

How is that supposed to work?

Gessmann: We are talking about digital agents here. Imagine a professional cyclist. He can ride a bike excellently, but he won't be able to describe exactly how he does it. It's the same with an excellent machine operator of a five-axis lathe.

There is a lot of experience and intuition involved. If we use AI to evaluate all operating parameters and machine data and transfer them into a digital twin - then valuable process knowledge could be preserved. But for now, that is more of a vision.

Speaking of visions: What about humanoid robots in the industry?

Gessmann: They are the logical next step after the cobot. Humanoid robots that move freely in a space and perform tasks - that is, of course, a huge opportunity for our living and working world and not just interesting in an industrial environment. But the real world and free movements in it are extremely varied and complex. In hospitals or manufacturing areas, the tasks and environmental conditions differ immensely.

For this, we need a precise simulation of the entire real world in the metaverse, where the digital twin of the robot can be trained. At Schunk, we have developed a humanoid robotic hand that mimics the human hand. We are currently working on bringing it to the highest simulation level five in the metaverse. In this way, we shape technological progress in all areas of automation and advocate for a healthier working world.