Setting the pace in the digital production age

Like Switzerland, Germany also has a "lively" robotics scene. Intelligent automation is finding its way into many areas. Automatica, the leading trade fair for intelligent automation and robotics, which takes place every two years, acts as a kind of "showcase". Due to Corona, the trade fair has been postponed; it is now scheduled to take place in Munich from December 8 to 11, 2020. The Fraunhofer IPA from Stuttgart will also be there to show what is already possible in terms of robotics and automation and where the journey on the shop floor of the future will lead.

Cooperative and networked navigation solutions

Compact mobile "rob@work" robots travel on an elevated exhibition area. They navigate autonomously, are networked with each other and show a miniaturized logistics scenario. Thanks to a continuous SLAM algorithm (SLAM = Simultaneous Localization and Mapping), the robots can reliably localize themselves even in changing environments without the need for additional infrastructure. In addition, they exchange data from their own sensors or from stationary sensors installed in the operating environment. This means that each robot always has an up-to-date map at its disposal, which it can use to adjust its route and localize itself. This avoids unnecessary routes, bottlenecks and downtimes. "With this cooperative navigation solution, we are demonstrating how driverless transport systems can enable matrix production, for example," explains Kai Pfeiffer, group manager of service robotics for industry and commerce at Fraunhofer IPA. "We can also add virtual robots to the exhibit and use augmented reality to visualize travel paths and other information," he adds. This simplifies and speeds up commissioning, maintenance or fleet expansions. The software has already successfully demonstrated the required agility of modern logistics processes several times in industrial applications.

Automated assembly and autonomous gripping

Many companies are concerned with the question of the extent to which they can automate their assembly tasks. For many years now, Fraunhofer IPA has been offering the automation potential analysis (APA) for this question. Until now, the APA was linked to the knowledge of an automation expert. A new app now makes this knowledge more easily accessible. It guides users to analyze their own assembly processes, evaluates their answers and provides information about automation potential. "With our app, anyone can become an expert in evaluating assembly processes," explains Alexander Neb, who works as a research assistant at Fraunhofer IPA and co-developed the app. It can be obtained via a simple license agreement for test use.

Another software for assembly automation is NeuroCAD. It analyzes component properties with the help of machine learning methods and uses them to determine the extent to which a component is suitable for assembly automation. Users can upload their STEP files free of charge at www.neurocad.de and find out within a few seconds how easy or difficult it is to separate a component. The tool also evaluates the gripping surfaces and alignability of the component. In addition, the neural network gives a probability that it is correct with its result.

Finally, the pitasc modular system for programming force-controlled assembly processes shows how manually executed processes can be automated in an economically viable manner. "Until now, it was necessary to largely reprogram a robot system for each application. With our software, tasks once modeled can be quickly transferred to new product variants, products and even robots from other manufacturers," says Frank Nägele, head of the Robot Programming and Control Group at Fraunhofer IPA. The software is structured similar to a modular system: It contains many ready-to-use and reusable program modules that can be individually assembled when setting up a robot system. pitasc is ready for use in pilot applications that the scientists would like to implement together with companies.

Not only assembly, but also the reach-in-the-box application is sometimes still a challenge for automation. With the exhibit "AI Picking", Fraunhofer IPA shows how machine learning methods and simulations significantly improve the application in terms of autonomy and performance. The scientists demonstrate this using the example of a robot that picks objects from an undefined position out of a box. For this purpose, an artificial intelligence (AI)-based object pose estimation provides robust and accurate object poses in a few milliseconds. "New objects can be taught quickly and easily based on a CAD model," explains project manager Felix Spenrath. "The software can also detect and solve entanglements and robustly handle packaging material." The robot was already extensively trained in the simulation and this knowledge was then transferred to the real application. Gripping poses are automatically generated and evaluated based on this knowledge.

Contamination-free production with protective cover and cleanroom tent

Not only more autonomous, but also ultra-clean production is increasingly in demand. "Clean production environments enable the high-tech of the future," explains Udo Gommel, head of the department of ultra-clean and micro-production at Fraunhofer IPA. "Tomorrow's key technologies will only advance with purity technology. It is crucial: from battery production to biotechnology." Fraunhofer IPA will use the following three products to demonstrate the technological advances:

• Protective sheath 2ndSCIN®: Freshly patented, 2ndSCIN® makes dynamic automation components such as a robot ready for ultra-clean production. The sheath consists of a permeable, movable and multi-layer textile, which is modeled on human skin in its mode of operation. Depending on the application, two or more layers can be superimposed. The layers are separated by spacers. In each gap, for example, air can be sucked in or removed. In this way, particles originating from the environment or from the automation component can be removed. The introduction of gases into the interstitial spaces of the system enables its sterilization. In addition, the cover can be changed in about an hour and can be reused after decontamination. The textile layers are also equipped with sensors that continuously measure parameters such as particle quantities, pressure or humidity. In the future, this sensor data will be evaluated with the help of CI algorithms and will enable predictive maintenance, for example.
• Mobile cleanroom CAPE®: Scientists from Fraunhofer IPA have also developed a mobile, tent-like cleanroom system that can be set up in less than an hour both indoors and in weather-protected outdoor areas. This "cleanroom on demand" provides manufacturers with a mobile, contamination-free manufacturing environment that enables air purity of ISO classes 1 to 9. The solution is suitable for manufacturers who need to produce contamination-free, but do not require a permanently available sterile and clean environment. Examples include applications in chip manufacturing, medical technology, the food industry and satellite assembly. The automotive industry also benefits from the cleanroom tent, for example in battery cell or fuel cell production.
• Fraunhofer Tested Device®: For many years now, the Fraunhofer IPA has also been offering methods for measuring particle emissions and awarding tested objects with the "Tested Device" certificate. In the aforementioned CAPE®, this procedure is demonstrated by means of an optical particle counter and a test object. With the product- and customer-specific test report, companies receive confirmation of the cleanliness and cleanroom suitability of their systems, devices or consumables.

Explaining machine learning and communicating data In robotics, as well as in numerous other fields of application in production and services, machine learning methods and artificial neural networks are increasingly being used. Depending on the application, it is becoming increasingly important to know exactly how they work and why they arrive at a certain result. They must become explainable. Due to their complexity, this is often not yet possible. "The more powerful a neural network, the harder it is to understand ", explains Prof. Marco Huber, who heads the Center for Cyber Cognitive Intelligence (CCI) and the Image and Signal Processing department at Fraunhofer IPA. Under the motto "Explainable AI" (xAI), Fraunhofer IPA presents methods that visualize neural network decisions and make them transparent and comprehensible for the user. "This comprehensibility strengthens the acceptance of AI, creates trust, improves correct functioning and provides legal certainty," explains Huber.

Data accumulates in every production, but it is often not possible to use and evaluate it due to different formats and interfaces. This is exactly where the "StationConnector" software comes in by providing a uniform interface across all plants. In this way, it can communicate data easily and application-specifically between industrial protocols, controllers and any IT systems. "With our software, users can quickly generate and implement data-based business models ", says Marcus Defranceski, group manager for purity-specific automation systems. The software is easy and flexible to use and is suitable for a wide range of applications, such as CI processes or monitoring.

Making production more efficient and relieving the burden on workers

A demonstrator for autonomous production optimization shows how losses in production can be automatically detected and their causes determined. It represents an automated model of a production line. This is observed both via the controller and via external sensors such as light barriers or cameras. All observation sources are used to create a behavior model of the line. This makes it possible to continuously analyze the line online and thus record normal behavior and identify production losses based on this. "In this way, we want to increase the effectiveness of the entire system and make key process parameters transparent," explains Julian Maier, scientist at Fraunhofer IPA and co-developer of the demonstrator.

Despite many automation possibilities, the flexible working power of humans in production is still irreplaceable in many places and must be preserved as best as possible. Exoskeletons, i.e. robot systems that are worn directly on the body, provide power support for strenuous activities and relieve the strain on humans. At the Fraunhofer IPA, there is the Stuttgart Exo-Jacket (SEJ), an exoskeleton for research and development purposes. The SEJ actively supports the upper extremities during lifting and overhead activities. The current system, demonstrated by Fraunhofer IPA, is mainly aimed at applications in logistics, where workers manually handle objects such as tires, boxes or suitcases with two hands in the area between knee and shoulder height in front of the body. "The core idea of the system is that users can still move their hands in the best possible way and thus make optimal use of their handling capabilities," says Christophe Maufroy, group manager of Physical Assistance Systems and Smart Sensors at Fraunhofer IPA, describing the special feature of the SEJ. The Stuttgart Exo-Jacket probably shows what could be understood by a "unity of man and machine" in the future ...