LÉO RAMOS CHAVES
Engineer Frank Allgöwer says that the lack of understanding of how complex interlinked systems work causes instabilityLÉO RAMOS CHAVESThe importance of basic research for the advancement of the knowledge and well being of humankind is a topic that interests engineer Frank Allgöwer, 54. And not just because he is the vice president of the German Research Foundation (DFG), the agency that funds projects that in most cases are related to basic research, with themes that are freely selected by the researchers. Allgöwer, a cybernetic engineering expert, argues that basic science is essential for solving specific and emerging technological problems. As an example, he mentions his line of research, the study of complex systems that operate in a network and are highly autonomous, and that support advances such as the development of autonomous vehicles and smart power distribution systems. According to Allgöwer, the director of the Institute for Systems Theory and Automatic Control at the University of Stuttgart in Germany, the lack of theoretical understanding of how interlinked systems work causes instability, observed in smart energy distribution grids. He believes it is necessary to develop mathematical models capable of predicting the operation of complex systems in order to control their performance.
Allgöwer spoke about these and other issues during a lecture in São Paulo on December 1, 2016. The event included the Leibniz Lecture, a DFG program to foster dialogue among the winners of the Leibniz Prize, one of the most important in German science and the scientific community. Allgöwer received the prize in 2004 for his studies on nonlinear systems and control theory. The presentation also marked the 10th anniversary of the cooperation agreement between the DFG and FAPESP, through which 14 research projects involving researchers from São Paulo and Germany have already been funded in various fields of knowledge. The DFG has already financed a total of 174 research projects in Brazil, almost half of which (86) are in the state of São Paulo.
In the German science and technology system, the DFG funds studies in which the scientist chooses the theme of the research, usually related to basic science. What impact does this type of research have?
In Germany, curiosity-driven research is seen as a force that generates knowledge that a few years down the line will produce innovations and have impacts on the industrial sector and society. This research has a high standard and involves significant amounts invested in basic science in Germany. While corporate research information is protected and caution is taken not to reveal sensitive data to competitors, basic research makes an effort to obtain group participation, with everyone doing research on interesting subjects. It works well in Germany.
What do you mean by the new cybernetics for the 21st century that’s the theme of your lecture in São Paulo?
It is very similar to the previous question on basic research. We are seeing complex networks emerge with growing frequency. They are cropping up in different fields, such as “Industry 4.0,” in which flexible production lines are based on machines that work together intelligently and manufacture various goods and commodities as inexpensively and productively as possible. Networks are being built everywhere. But even now, little is understood about how they work and the basic underlying principles behind complex networks. The result is that, right now, they are not entirely secure. Take smart energy grids for example. In the past, energy was generated at large power plants. Today, more and more renewable energy is being produced in small units, and distribution is decentralized in smart grids. In some parts of the world, including the United States, there have been blackouts, knocking out power to large areas. The problem is not the power plant or a transmission line but rather how the grids operate. No one can figure out what causes these problems. New cybernetics attempts to help people understand these things by performing basic research and investigating why something works or doesn’t work. The goal is to be certain that interlinked systems will operate properly. It is important for basic research to follow the requirements of existing technologies.
James McLurkin
Robot swarms, machines that work together and interact, are increasingly less costlyJames McLurkinIn your opinion, what is the future of smart energy distribution systems?
I will use Germany as an example, where we have a situation that is anything but typical. The renewable energy in southern Germany is generated on wind farms in the north. It has to be moved from one place to another. This is done using decentralized distribution systems as opposed to large transmission lines. Substantial amounts of energy are generated when winds are strong, but when they are calm, renewable sources from other places are used. The energy may be solar or hydroelectric, and it moves over smart grids. It is very important for us to have smart power distribution systems to make adequate use of renewable energy. You can see how smart grids will be important in developing countries. I believe that basic research, and not just applied research and innovation, can make a contribution to this development. We have to be certain that there will be no catastrophic events in the operation of grid systems that make them unstable or cause losses of power generation.
Could you explain the “nonlinear predictive control method” that you use? What are its applications?
Computational algorithms are already being used to control systems in a variety of fields. We want to understand what is working and not working and when. We want to find solutions to ensure that everything runs smoothly. The term “control” has to do with systems over which you want to exercise influence so that they perform in a given manner. Predictive control is a special method built upon the idea that you can have a mathematical model of the system you want to control and then predict how it will perform, assuming we apply some stimulus to it. Predictive refers to the fact that the model attempts to evaluate what will happen. Nonlinearity is a mathematical term that explains unexpected behaviors of complex systems; if you double the amount of stimulus, it doesn’t mean that the response will be twice as great, and it may produce something completely different. For example, nature is typically nonlinear.
To what extent does the way nature works inspire research on complex networks?
My group focuses more on the technological side, but the basic principles, math and methodology are the same for the groups that study these things from nature’s point of view. We can observe nature and learn to solve problems using it. One example in particular is birds flying in flocks. To fend off attacks by eagles, they form groups and confront the predator like a herd of animals. They don’t worry much about where they are flying to; they stay together and manage to avoid colliding with each other. This was a research theme several decades ago: how are birds able to do this? Do they look at the bird next to them? Do they look at the two birds next to them? Do they look in one direction only and not see what’s behind them? How do they observe the environment so that they can fly perfectly in the air without flying into each other? The discoveries are already being used with mobile robots. Birds look at five nearby birds at the same time. Each of the animals in the group is looking at exactly five birds. It doesn’t matter whether there are a few dozen or a few hundred birds in the flock. Each bird looks at five other nearby birds, and that’s all it takes for the system to work like a charm.
Is this related to research on swarm robots, small machines that work together and are integrated?
That’s not exactly my field. I’m a theoretician and I develop methods to research swarm robotics. But I do know that quite a bit is happening. One of the things that has changed in the last 10 years is that today it is much less costly to build small robots that work in groups. In the past, research institutions that had projects in this field asked for substantial funding for agencies such as the DFG or FAPESP because they needed a hundred or so small robots, and each one was very expensive. They are much less expensive today. This is a very active area of research in Germany.
What technological challenges are involved in the development of “Industry 4.0?”
Technological changes have already taken place. We have systems in decentralized bases and advanced cognitive skills. It is also possible to produce large amounts of data and disseminate it. Having communication systems that work together is within our reach. Now specific uses have to be developed. That is the challenge for Industry 4.0. It’s not about progress in terms of developing sensors or any technology in particular; it’s about obtaining a deeper understanding, in addition to creating scenarios for application and standardization. One example is getting robots from different companies to easily talk to each other—speaking the same language.
Tony Webster/Flickr
Wind energy generated in northern Germany is distributed to southern Germany through decentralized smart gridsTony Webster/FlickrDo you think that developing countries such as Brazil are ready for this change?
Brazil is a country that places heavy emphasis on manufacturing technologies, as is also the case in China and India. To my knowledge, Brazil has done quite a bit of work on Industry 4.0, including projects cofinanced by the DFG. Brazilian and German researchers are collaborating in this area.
When will we see autonomous vehicles on the streets of big cities?
I think there will be two different types of development and we need to distinguish between them. One is for vehicles such as trucks and buses that can move about autonomously on certain types of streets, such as highways that connect major cities, for example. I think we are very close to having trucks and buses that run autonomously on these highways. The other is to have autonomous cars on the streets of an interesting city like São Paulo in the midst of pedestrians and congestion on some thoroughfares. That’s a challenge. It will still take some time before we have a significant number of autonomous cars under these conditions. But the problem is more than just technology. There are legal issues as well. Who is liable when autonomous cars collide? Is it the manufacturer? Is it the owner? Or is it the programmer who developed the codes used by the cars? This will have to be settled before we see large numbers of autonomous cars on the streets.
How do you feel about the 10 years of partnership between the DFG and FAPESP in which 14 projects have been funded?
As a research funding agency, evaluation is our forte, but I think it would be inappropriate to look only at the numbers and conclude that the more projects that are cofinanced, the better the partnership is. The important thing is the quality of the cooperation and the opportunity to have good researchers from the two countries work together on important projects. In this regard, we are very satisfied with the cooperation agreement. The DFG and FAPESP think and work in much the same way and joint projects are carried out seamlessly.
What similarities are you referring to?
My understanding is that FAPESP, like the DFG, works very independently. We do not follow government orders as to which projects to fund. Obviously, the focus is on quality: the quality of the research, the projects, and the researchers. This is key. There are other similarities. In FAPESP and the DFG, researchers can submit projects at any time. In some agencies there are one or two deadlines during the year for submitting projects, but not at the DFG. With similar models, it’s easy for the two institutions to work together.
The DFG cooperates with other countries in Latin America, such as Argentina, Chile, Mexico and Colombia. How significant is the agreement with FAPESP in the lineup of DFG projects in the region?
This agreement is our standard. The only city in Latin America where the DFG has an office is São Paulo. In relations with Brazil and the state of São Paulo, FAPESP in particular is one of our key contacts.
