Munich For a long time, robot manufacturers have been promising industrial customers simpler operation. However, commissioning and programming the mechanical helpers is still complex and time-consuming. However, artificial intelligence (AI) could now ensure real progress. “AI can become a real game changer for the robotics industry,” predicts Frank Konrad, CEO of Hahn Automation and chairman of VDMA Robotics + Automation.
An example: the Munich start-up Fruitcore Robotics, which is presenting an AI-based operating system called horstOS at the Automatica trade fair. The software should be easy to use. HorstOS is intended to support users in all important steps, from programming the tasks to the concrete management of robot components such as cameras and gripping systems. The program also supports integration into the production processes.
Fruitcore means “easy as pie” quite literally: last week, a group of schoolchildren were able to test the technology, which is now to be used in all of the start-up’s robots. The children programmed the waypoints of a cardboard labyrinth and transported chocolate bars from one box to the other.
Robots should be programmed using colloquial language
An AI co-pilot makes this possible. This can communicate with users, give tips for troubleshooting, but also write entire programs themselves. It can be used, for example, with the new Horst1000 robot, which can take over the loading and unloading of machines, among other things.
“Programming with slang has always been our goal,” says Fruitcore CEO Jens Riegger. Therefore, one was well prepared when language models made the breakthrough with ChatGPT last autumn.
With the help of the AI co-pilot, you can now tell the robot: “Drive to the right corner of the tabletop” or “Lift the glass ten centimeters”. Previously, this required complex programming. “The use of robotics will become affordable and, above all, accessible,” says Riegger.
The providers of the so-called collaborative robots have been able to show good growth rates in recent years. However, the big breakthrough that many had hoped for did not materialize. “Cobots” are still the exception, especially in many smaller companies and in the trades.
>> Read here: Seven graphics show where in the world the development of artificial intelligence is concentrated
However, delivery difficulties during the pandemic and the Ukraine war as well as the shortage of skilled workers have triggered a rethink in companies. “There is currently a push for the topic, we can feel it,” says Danny Denk, CEO of the integrator Ecosphere Automation.
With artificial intelligence, the machines should learn to feel haptically
However, the “democratization of robotics” will only succeed if there is something “hands on” because many have never come into contact with a robot. That’s why he opened a “Solutions Experience Center” where the robots can show what they’re already capable of.
Denk is also convinced: “The robot needs a brain.” On the one hand, AI can help with programming. In addition, the machines could now learn to recognize objects haptically and optically more quickly. So far, for example, robots have found it difficult to alternately grip very different objects.
“Here you can now make progress with AI,” says Kuka boss Peter Mohnen. The largest German robot manufacturer uses the technology for “Smart Bin Picking”. The machines are constantly learning, for example how to grip glasses without breaking them.
The managing directors of Fruitcore Robotics see enormous potential in their approach, even for small companies.
(Photo: Fruitcore Robotics)
Cobot world market leader Universal Robots is also presenting Polyscope X at the Automatica trade fair, software that uses the first AI elements. “Artificial intelligence offers great opportunities,” said UR Vice President Anders Billeso Beck to the Handelsblatt. For example, it could help to increase the reliability of production. So far, the systems have often been stopped automatically in the event of the smallest deviation – for example because a part has a slightly different shape. AI can help to deal better with variations.
>> Read here: Up to 300,000 euros salary: Five job profiles created by AI
UR recently registered a reluctance to invest among customers in view of the difficult macroeconomic environment. But Beck is also convinced that the shortage of skilled workers will be the decisive driver. Ten years ago, potential customers would have been interested in the return on investment: they wanted to know whether a robot would quickly pay for itself. Today they wanted a solution so that they can produce at all: “Automation is often the only solution to this problem.”
Other robotics specialists are also working intensively on AI integration. The Karlsruhe company Artiminds is developing software with which the use of robots can be planned and programmed. This allows the machines to decide for themselves how much to accelerate or how hard to grip when they are grinding something or gripping irregularly shaped objects. The system is constantly learning and optimizing itself during operation, says co-founder Sven Schmidt-Rohr.
The goal is three-digit million sales
Fruitcore believes it has an edge over the competition. “We want to offer the most intelligent robots,” says Managing Director Patrick Heimburger. GPT is only the tool: “We train the models with our own data.” The entry barriers for competitors are therefore high, especially since Fruitcore benefits from the combination of hardware and software.
>> Read here: Machine builders warn: “China automates faster than Germany”
There are also skeptical voices in the industry. One has to wait and see whether it is more than a gimmick and whether operation with everyday language is necessary and helpful for professional use in industry, says an expert.
However, the company was able to convince a number of customers and investors. At the end of last year, Fruitcore raised 23 million euros in a new financing round from venture capitalists such as UVC Partners and Capricorn. A good 800 robots have been installed at customers so far.
This text is part of the large Handelsblatt special on artificial intelligence. Are you interested in this topic? All texts that have already appeared as part of our theme week
You will find here.
With the proceeds from the financing, the Horst models are now to be rolled out widely. The technology is also suitable for large robots, says Riegger. “In the medium term we want to achieve three-digit million sales.” Fruitcore Robotics wants to further develop the technology itself – and is therefore not striving for integration into one of the large corporations. “We’re not tired enough to sell ourselves yet.”
With the help of artificial intelligence, the degree of automation in factories in Germany could increase over the next few years, and domestic robotics specialists could play a leading role. However, Konrad from the VDMA sees one prerequisite for this: “We have to be careful not to dead-regulate artificial intelligence.”
More: German robotics and automation industry expects a record year.