DE

Practical Relevance of Future Technologies

Intelligent assistance systems elevate human-technology interaction to a higher level in automated production. Institute Director Dr. Matthias Peissner of the Fraunhofer Institute for Industrial Engineering and Organization IAO assesses future technologies for workplace 4.0 for their practical relevance.

In your view, what does the digital workplace of the future, “Workplace 4.0”, look like? And what distinguishes it from current workplaces?

Dr. Peissner: There is no such thing as an intelligent workplace. The idea of the smart factory is based on the intelligent networking of people and machines. This Industry 4.0 scenario also enables new forms of human-machine interaction. In general, fields of application for digital assistance systems and their design are becoming clear through the analysis of the interaction between processes and workforce. Our aim at IAO is to make human-technology interaction more natural and casual.

What does that mean specifically?

Dr. Peissner: Employees focus on the content-related technical tasks, and interaction with the system simply runs along. For the design of intelligent assistance systems, there is a kit with technologies such as multi-touch recognition, smart watches, headsets for voice interface and head-mounted displays for augmented reality, where the image of the real world is superimposed with digital information. In addition, many assistance systems are only possible due to artificial intelligence. Machine learning methods are the basis for decision support systems and also for many detection technologies.

Human-Technology Interaction in Workplace 4.0: Control as on a Smartphone

You want to make human-technology interaction more natural and casual. Does this include wiping as well as gesture and voice control, as we know them from a smartphone?

Dr. Peissner: Yes, of course For example, Fraunhofer developed the “Documation” system for quality assurance with BMW, which facilitates documentation when inspecting painted surfaces. If the test part is OK, the tester wipes it from left to right. If he points his finger at a paint defect, his gesture is automatically detected and the position on the bumper is recorded in the system. Information can also be added via speech. This saves a lot of time in actual practice. Acceptance is high because documentation becomes easier, and the number of neatly documented cases also increases.

In addition to everyday technologies such as wiping and speech, does your research area for Workplace 4.0 also experiment with completely futuristic concepts?

Dr. Peissner: A spectacular example is certainly our Brain Computer interface for production, which is based on electroencephalography or EEG, i.e., measurement of the electrical activities of the brain. This enables mental states to be used for interaction. If a user is presented with workpieces on a conveyor belt during a quality test and his EEG signals are measured, he can automatically sort out faulty workpieces via the Brain Computer interface.

Workplace 4.0 Assistance Systems – also a reduction of the burden on blue collar workers

You said that many assistance systems are only possible due to artificial intelligence. What kind of assistance systems are relevant for low-skilled workers, for example?

Dr. Peissner: Guidance, error prevention and qualification systems are particularly relevant for low-skilled workers. These can specify context-sensitive processes step by step and provide feedback for readjustment in the event of incorrectly executed procedures. Assistance systems can also make work steps natural and intuitive. Examples are gesture-controlled robot programming or the ability of an employee to set up his own tools.

In which application scenarios is such a system worthwhile?

Dr. Peissner: A Fraunhofer exhibit for the future of collaboration with robots outlines our vision very well. Today, the use of robots is primarily profitable in mass production, because setting up robot systems is time-consuming and requires proven expertise for programming. With the gesture-based human-robot interface Insitu, we enable blue collar workers, i.e., the people on the production line, to produce a robot program or to adapt it to the current situation. In doing so, they specify certain parameters of the program through gestures that are recorded by high-performance and inexpensive 3D and color cameras and interpreted by a system.

How can an assistance system guide the production line and what is the challenge in design?

Dr. Peissner: In a current project with a large automotive component supplier, we have a pilot system that supports an employee on a semi-automated production line who handles several machines. Up to now, depending on the process in progress, he switches from machine to machine based on his experience to add material, start a sub-program or correct an error. Our algorithm then calculates optimum work processes based on the current machine and sensor data, and the employee is informed by LED strips on the floor, via a smartwatch or a large display system in the hall where he should optimize the process next.

Doesn't the employee feel patronized?

Dr. Peissner: The challenge is not to lead employees by the nose through such AI-based systems; they should feel supported to make the most of their manpower. Otherwise, frustration, stress, fatigue, overburdening and illness are pre-programmed. Only human-friendly assistance and automation have a future.

More Efficient Processes thanks to Assistance Systems in the Smart Factory

What support can assistance systems at Workplace 4.0 provide to the higher-skilled workers, such as maintenance engineers?

Dr. Peissner: This involves solving complex problems or finding causes of errors, for example. A chatbot provides support here, i.e., a dialog system that queries in a structured way in case of a problem and restricts the search space. The media that display information from knowledge management can be selected according to requirements, for example tablet, smartphone, smart glasses or AR system. Fraunhofer also has a demonstrator that automatically converts machine faults into tasks for appropriately trained maintenance personnel. As a result, a problem only ends up in the cockpit of those experts who have the necessary skills and experience.

And how is a person responsible for the overall process helped in the Industry 4.0 context?

Dr. Peissner: A shift supervisor receives decision support. He wants to know how to adjust process parameters in terms of efficiency – should it be in terms of quality or speed? Such systems require digital models of the processes, which represent a realistic representation of the production process. Unfortunately, they are still rarely completely available in practice. But when they are, I can use parameters for a new order to forecast energy consumption or the quality of production results. The result is information based on current process parameters and forecasts for the optimization of parameters in accordance with current requirements. This enables employees to make decisions competently even if they do not have a complete overview.

These five trends for Workplace 4.0 should be kept in mind

The selection and design of the appropriate intelligent assistance system is highly dependent on the workplace. Dr. Matthias Peissner, Director of the Institute and Head of the Human-Technical Interaction Research Unit at the Fraunhofer Institute for Industrial Engineering and Organization IAO, advises keeping an eye on five trends and technologies.

  • Machine vision is used to identify objects in production and to detect their position direction. “This can be used to automate quality assurance tasks. We can see if the right part is being used.”
  • Acoustic analyses can detect irregularities in operating noise. “They are suitable for monitoring the condition of machines, among other things.”
  • Mixed reality technologies have potential in prototyping or maintenance. “Maintenance is facilitated, for example, by the superimposition of real and virtual data via augmented reality.”
  • Chatbots are technical dialog systems, with which text input or speech is communicated. "A maintenance engineer can interact elegantly with an expert system via a chatbot, analyze problems or perform complex test tasks in a guided manner.”
  • Methods from machine learning are used for decision support at the process level.