Table of Content
Research Overview
In one way or another, most of Ralf’s research activities can be attributed as embedded systems, in a sense that they are small, rather custom-made devices that directly interact with their environments. The research activities comprise, for example, high-precision localization, neural implants, self-organizing smart appliances, ambient assisted living and health care, autonomous mobile robots, secure application-level gateways, hardware raytracing, and sensor networks.
In these systems, the following methods are of particular interest: cost-efficient hardware design based on field-programmable gate arrays (FPGAs), soft computing methods, such as evolutionary algorithms, neural networks, and fuzzy logic, new artificial intelligence, and last but not least common sense. For a more comprehensive overview, please visit the Projects Page.
In addition, several of our research projects have received some awards and appreciations as can be seen on the Awards Page.
Research Strategy
Ralf strongly believes in inter- and transdisciplinary research, and thus prefers heterogeneous working groups. Because of his stay at Rolf Pfeifer’s AI Lab at the Institute of Computer Science Ralf’s research draws significant inspiration form biological observations. The idea behind this is that nature has had millions of years to develop optimal solutions for a huge number of problems, which might also be relevant for current research efforts.
Quite a good example is the transfer of the barn owl localization system into a working technical implementation. Our first prototype, called X-ORCA, has been implemented on a standard, low-cost FPGA and achieves a time resolution of about 10ps, which corresponds to a traveling distance of about 3mm of an electromagnetic signal.
Furthermore, Ralf likes research on autonomously behaving systems, i.e., systems that mainly operate without any human intervention, since these systems naturally raise new research questions and since they are often moving, which is just more fun. In order to bring this view to students, Ralf is engaged into integrating the e-puck robots as well as a Model Railroad Project into regular teaching.
Research Methods
The concept of embedded systems is widely applicable, and thus comes in various flavors. Ralf’s research aims to "build" rather custom-made systems that exhibit an autonomous behavior to a certain extent. Since they have to be custom-made and/or have to be realized as prototypical systems or just as proof-of-concepts, most hardware-oriented research is based on low-cost, general-purpose field-programmable gate arrays (FPGAs). Even though FPGAs are generally much slower than most off-the-shelf processors, they can exploit parallel processing by their very nature. In order to obtain a good compromise between an performance and (hardware) costs, the concept of hardware-software co-design is also highly relevant in most cases.
Because most of the developed systems have to exhibit a some autonomy, which more or less implies adaptation and/or learning, Ralf’s research draws significant inspiration from biological systems. It thus does not come to a surprise that he has been doing significant research in the areas of soft computing methods , particularly evolutionary optimization and neural networks . In addition to these methods, Ralf also draws significant inspiration from actual biological systems, such as crickets, ants, and barn owls.
Collaborations
Ralf maintains active collaborations with other national and international research institutes. This includes Prof. Dr. Sven Bruhn at the University of Rostock, Dipl.-Ing. Gerald Bieber at Fraunhofer IGD in Rostock, Prof. Dr. Dirk Arnold at Dalhouse University, the Institute of Neuroinformatics at the University of Zurich, Prof. Dr. Christopher Lueg at the University of Tasmania, Prof. Dr. Auke Ijspeert at EPFL (Lausanne, Switzerland), Dr. David Fogel at Natural Selection, Inc., Sebastian Staamann, and Prof. Dr. Stefan Rüger at Imperial College London, to mention but a few.