Biointelligent Systems

“NATURE NEVER THOUGHT OF INTRODUCING A CENTRAL COMPUTER.”

 

Fraunhofer IML and TU Dortmund University created a unique test environment for the human-machine-interaction with the “Innovation Lab Hybrid Services in Logistics”. In an interview with “Logistik entdecken”, Prof. Dr. Dr. h. c. Michael ten Hompel, Managing Director of Fraunhofer IML, talks about the work of the future, artificial intelligence and biointelligent systems - and explains what a swarm of drones has to do with all this.

 

What does “humane work” or “responsible digitization of work” mean to you in this context?

 

Responsible action is acting according to accepted standards. These can be laws or societal standards. The term “responsibility” was coined for human beings. As of today, only humans can accept responsibility. However, people are confronted with more and more intelligent machines that are able to adapt their behavior and to learn. In this context, we created the term “machine responsibility” - actually a contradiction in terms. It points towards a future in which people and machines will organize responsible action collaboratively. In addition, we live in a globalized world. In global systems, we cannot assume that all actors share our social and Christian world view. In China or even in the United States, we are confronted with other standards to which we have to adapt our systems. For this reason, we have changed the term “responsibility” to “machine responsibility”.

 

What will a typical job in logistics look like in ten years?

 

At first glance, the workplace will not look any different at all in operative logistics. At second glance, however, you will notice that the employees move about in an intelligent environment, that they talk to their environment and that it changes dynamically. Processes will be introduced that relieve people in their work, both cognitively and physically. Active and passive exoskeletons, for example, will be in use - some of them are already used today. People are increasingly becoming conductors and organizers who organize work in cooperation with machines and do less and less physical work. I do not think that the production halls will be empty of people in ten years, but the workplaces will look completely different.

What does “humane work” or “responsible digitization of work” mean to you in this context?

 

Responsible action is acting according to accepted standards. These can be laws or societal standards. The term “responsibility” was coined for human beings. As of today, only humans can accept responsibility. However, people are confronted with more and more intelligent machines that are able to adapt their behavior and to learn. In this context, we created the term “machine responsibility” - actually a contradiction in terms. It points towards a future in which people and machines will organize responsible action collaboratively. In addition, we live in a globalized world. In global systems, we cannot assume that all actors share our social and Christian world view. In China or even in the USA, we are confronted with other standards to which we have to adapt our systems. For this reason, we have changed the term “responsibility” to “machine responsibility”.

 

In July 2018, Fraunhofer IML opened the “Innovation Lab Hybrid Services in Logistics” together with university partners - funded with 10 million euros from the Federal Ministry of Education and Research (BMBF). What can the laboratory contribute to this vision of workplaces of the future?

 

Meanwhile, sociological topics and the design of workplaces have a completely different meaning. Within the Innovation Lab, we intend to work on a vision of the future that we call Social Networked Industry. And with this term we mean, on the one hand, the social design of work and, on the other hand, we assume that cooperation between people and machines will take place in social networks in the future. This creates a completely new relationship between humans and machines.

“We need artificial intelligence techniques to realize the partnership between man and machine.”

And this relationship needs to be investigated sociologically.

 

Correct. However, for this vision of the future we need a lot of technology that we also develop in the innovation lab. These technologies include an autonomous swarm of drones that serve as test vehicles for simulating fast autonomous systems.

 

Why is a swarm of drones needed to do that?

 

We use the swarm of drones to develop algorithms with which we control the interaction with autonomous vehicles and machines as well as the interaction with humans. We are increasingly using artificial intelligence methods e.g. machine learning and neural networks. Drones are well-suited because we can virtually illustrate any scenario on an industrial scale in a three-dimensional and highly dynamical way. Controlling such a swarm of autonomous drones is also a research object in itself –  from sensors to real-time networking to AI algorithms for decentralized swarm organization. In addition, we also use ordinary vehicles and hundreds of intelligent containers that control the scenario. Our research lab is therefore just as intelligent an environment as it will be commonplace in future.

Artificial intelligence is on everyone's lips in times of Industry 4.0. Why do you prefer to talk about biointelligent systems? And what part will these systems play in the work of the future?

 

Artificial intelligence  does not exclude biointelligent systems, quite the opposite. We see that processes of artificial intelligence will play a very important part. You need artificial intelligence processes in order to realize the partnership between humans and machines aiming at creating efficient and humane systems in which the future of logistics can take place. Artificial intelligence, however, is becoming more and more applicable in relatively small cyberphysical systems from shelves to drones. The smallest current example of artificial intelligence is our Low Cost Tracker, which we developed together with Deutsche Telekom: flat rate for 10 euros, target price about 10 euros and communication via  “NarrowBand IoT” (NBIoT). Today, such systems have enough computing power to implement simple artificial intelligence methods. In future, we will have to deal with a large number of systems that are potentially capable of executing AI algorithms. Artificial intelligence then consists of individual decentralized cells. Biology would speak of protozoa.

Which brings us to the biointelligent systems.

 

Exactly. The question is how to control such systems? We have to take a look at evolution. It is the only thing that has survived the last three billion years. A lesson to be learned from that: evolution has never thought of introducing a central computer for a swarm, habitat or biosphere which is quite contrary to what we are currently doing. In future, however, we are talking about hundreds of thousands, millions or billions of such systems, all of which have the ability to act autonomously and on which weak and eventually strong AI algorithms will run. Another aspect is that we will be dealing with real-time networks in future (think of 5G). NB-IoT is an ultra-low power network which can only transfer small quantities of data and is not real-time capable.  

This will evolve within 5G, so that we will have the same real-time network for both intralogistics and extra logistics. But what about the virtual environment for artificial intelligence, the biosphere of our protozoon? In my point of view, this is the central research question for the coming years.

What do you think the environment will look like?

 

This is difficult to predict. The environment will definitely contain elements that we find on social networks on the Internet today. This biosphere will also be a further development of the Internet of Things. In the Internet of Things, things control themselves, they are simply preprogrammed accordingly. In future, however, we will be dealing with truly autonomous and self-sufficient systems. If we combine these with artificial intelligence, we will not only be dealing with learning protozoa but with learning overall systems eventually.

Back to the Innovation Lab: being a company representative, what can I learn in the Innovation Lab for the future of my company and my current challenges?

 

For the first time in human history, we are currently in a situation where we hold more technology in our hands than we can sensibly use. A new world of tremendous possibilities is opening up through digitization and computer technology, which is nowadays available for just a few cents. Let us just take a look at how little we use the computing power of our smartphones: we use it to run a few  apps and swiping the screen – and that with a computing power that would have been equivalent to that of a supercomputer in the year 2000. We help companies take advantage of existing opportunities to develop new technologies and business models. This is an essential task for us, also with regard to the cooperation of humans and machines. We provide the research environment to test things. Above all, however, our colleagues are prepared to share their knowledge with industrial companies and work together on projects and innovations.

To which extent does Fraunhofer IML benefit from the exchange with companies for its research agenda in the innovation lab?

 

The development of real artificial intelligence will only succeed in interaction with the application. Making a robotic system truly intelligent requires that the system cooperates with its environment, perceives the environment, cooperates with other robots and humans - and therefore learns and develops further. In an industrial environment, artificial intelligence processes cannot only take place in the computer itself but require interaction with the industrial environment. And these are exactly the projects with which industry is approaching us and vice versa because we push forward developments around artificial intelligence, machine learning, digitization, industry 4.0 and biointelligent value creation. In the end, the race for artificial intelligence will not be decided in the laboratory but on the hall floor.

Prof. ten Hompel, thank you very much for the interview!