The system that sweeps the Go world champion off the pitch. The car that drives itself through rush-hour traffic. The robot that greets visitors to a trade fair stand in a friendly manner. When the conversation turns to the topic of artificial intelligence (AI), these striking examples quickly overshadow the discussion. Yet AI is changing the world behind the scenes on many levels. The effects are hard to grasp: different work processes, business models and platforms everywhere. The impact of current developments becomes clear when we look at the use of AI technologies in physical objects. For every machine, every household appliance and every car with AI components, new possibilities are opening up, new processes are conceivable and new opportunities for monetization are emerging.

Cyber Physical Systems (CPS) are of particular importance here. They close the gap between the worlds by directly integrating physical objects and processes into digital processes. This increases the efficiency of existing processes as well as the accuracy and timeliness of the data processed in them. AI opens up new approaches to interpreting the physical world and the interaction between the system and reality. Typical scenarios in which AI technologies in CPS come into their own: recognizing patterns in complex processes or working closely with people on joint tasks, for example in construction or mechanical engineering.

New technologies - new possibilities

It feels like there are as many definitions of AI as there are researchers working in this field. A short, pragmatic definition will suffice for the following explanations: AI is a branch of computer science that deals with research into the mechanisms of intelligent human behavior. When applications translate spoken language into instructions, when machines recognize the level of fatigue of a human worker or an image recognition system marks objects on photos, AI technologies are working in the background.

It is precisely these technologies that are suitable for use in the CPS environment. This is because the scenarios described above, in which CPS demonstrate their strengths, are typical fields of application for AI.

The use of robots or other automated production systems was previously limited to executing predefined motion sequences. Machine learning (ML) technologies allow for much more flexible use in the production process. Thanks to ML, AI experts have achieved breakthroughs in the field of machine vision and hearing within a short space of time. This includes the automatic recognition of objects and situations, the interpretation of gestures and the understanding of spoken instructions. A typical application scenario: during operation, employees give commands to the system by shouting or using simple gestures. They do not have to interrupt their own work to do so. This allows production systems and the human workforce to work together more closely and smoothly.

CPS on its own

Another example of the potential of AI in an industrial environment is the ability to work autonomously. AI helps systems to decide and work more independently. For example, the methods of the aforementioned ML allow a CPS to automatically derive rules and develop adaptive behavior. The systems recognize connections or correlations and derive rules from them without experts having to define them in advance.

Such an intelligent and networked CPS independently estimates the effects of a delayed delivery of raw materials, for example. The system independently adapts the production process to the new delivery dates. It informs downstream or upstream production stages, reprioritizes the processes and allocates adjusted time slots - independently of human intervention. Thanks to AI technologies, the CPS does not rigidly follow a predefined program. It adapts its planning independently to the changing environmental parameters.

AI also opens up the field of predictive maintenance, understood as the early detection or prediction of problems that occur and the resulting predictive maintenance work. This is based on the concepts of pattern recognition and anomaly detection. AI offers the opportunity to avoid or at least reduce downtime. This allows companies to save costs and provide more reliable services. For critical systems, this not only leads to financial savings, but also to greater security.

AI technologies are not only opening the door to new use cases at CPS and production process level. Employees with strategic tasks also benefit.

Know more faster

New studies, new data, new framework conditions: Information flows into a company in a constant stream. Experts sift through, check, analyze and develop recommendations for action. One of the core tasks of so-called knowledge workers is to master extensive information and quickly gain important insights. It is almost impossible for individuals and teams to grasp all the relevant material and assess its significance.

One example in the industrial environment is the development of commodity prices. Pricing is a complex mechanism that depends on an unmanageable number of factors. For example, from legal requirements to geopolitical developments, from consumer habits to the outcome of elections in individual regions. Forecasts are correspondingly difficult. This is where AI solutions come into their own. On the one hand, the aforementioned pattern recognition makes it possible to recognize correlations that remain hidden to human observers due to their complexity.

On the other hand, AI technologies help experts to stay up to date in their subject areas. This is because solutions can now deal well with texts written in natural language. So-called natural language processing (NLP) techniques enable them to grasp the meaning of content, summarize documents and present these summaries in the form of dashboards, for example. With the help of machine learning methods, the applications extract the topic and information about the semantics and summarize the central content.

Whether a study, position paper, legal text or scientific treatise: trained AIs prepare the core statements in such a way that experts can quickly gain an overview. If required, they can then access the sources of the information. Those responsible set up fine-tuned reporting systems that provide them with targeted, automatically prepared dossiers on individually relevant topics.

Designing the production chain better

In the factory of the future, better informed people will work more closely with more autonomous machines. This ensures that those involved can organize the entire production chain differently. From the initial planning of a product to its construction, delivery and operation: AI applications will help companies achieve their goals with fewer resources throughout the entire process.

Prof. Dr. Volker Gruhn, Chairman of the Supervisory Board and founder of Adesso AG © Adesso

Prof. Dr. Volker Gruhn, Chairman of the Supervisory Board and founder of Adesso AG