Research

The Chair for Design Automation covers the development of design methods for various application areas. A particular focus of our work is on the design, verification, and test of circuits and systems for conventional as well as alternative and post-CMOS computing technologies. Besides that, we have proven to successfully apply the methods developed by us in complementary research areas. In the following, a summary of our research interests and contributions is provided. Besides that, you can find our publications on this page.

Please also note that dedicated pages exist which summarize our work on

• Quantum Computation,
• Microfluidic Biochips,
• Field-coupled Nanocomputing,
• Machine Learning, and
• Design Automation for the European Train Control System.

Design for Conventional Computing Technologies

The design of today’s computing devices (including embedded and cyber-physical systems) is one of the most complex problems Electronic Design Automation (EDA) is currently facing. In order to handle the ever increasing complexity, designers constantly introduce higher levels of abstraction. Today, design at the Register Transfer Level (RTL) and the Electronic System Level (ESL) is common. In our work, we are developing algorithms which improve existing design and verification techniques for these abstraction levels. At the same time, new abstraction levels are considered which lift the design process to abstractions closer to the initial specification provided in natural language. For this purpose, modeling languages such as UML or SysML as well as techniques from Natural Language Processing (NLP) are exploited. Within this area, we consider the design from the initially given (textual) specification to its first (formal) representation provided in terms of UML/OCL, SysML, MARTE, etc. This led to contributions in domains such as

• Mapping of natural language specifications to formal models,
• Verification and debugging of formal models provided in UML or SysML,
• Modeling and implementation of non-functional behavior such as timing, and
• Generic representation of functional and non-functional behavior from different description means.

Next, how to guarantee a (correct) realization of the resulting model e.g. in terms of a system implementation or circuit netlist has been considered by us. This resulted in contributions for

• Verification and debugging of designs implemented at the RTL and the ESL and
• Design understanding and visualization of circuits and systems implemented at the RTL and the ESL.

In addition to that, we also consider research questions at lower abstraction levels. Here, physical issues (represented by fault models or models for energy consumption) have to be considered in addition to the purely functional description. More precisely, we further contribute to the fields of

• Automatic test pattern generation at the gate level,
• cost estimation,
• place and route as well as floorplanning, and
• the design of low power interconnect encoders.

In these areas, we are also more and more use methods for Machine Learning, Reinforcement Learning, etc.

Design for Alternative and post-CMOS Computing Technologies

While the previous decades have witnessed impressive developments in the design and realization of conventional computing devices (predominantly based on CMOS), physical boundaries and cost restrictions led to an increasing interest in alternatives. Quantum computing, reversible computing, microfluidic biochips, field-coupled nanotechnologies, optical computing, memristors, DNA computing, and further alternatives are being discussed at the moment. Besides the physical, biological, or chemical aspects, these emerging technologies also require thorough basic research on how to efficiently design these future circuits and systems. In our work, we are performing research towards proper design flows for alternative circuit and system technologies. This particularly includes works on quantum computation, microfluidics, and field-coupled nanotechnologies.

For quantum computation, we develop methods for

• Efficient representation and manipulation,
• Simulation of quantum circuits,
• Synthesis of quantum (and reversible) circuits,
• Mapping of quantum circuits to real architectures, and
• Verification of quantum computations

For more details on our work in the domain of quantum computing, please see this page.

For microfluidics, we develop methods for

• Simulation and
• Design automation

For more details on our work in the domain of microfluidics, please see this page.

For field-coupled nanotechnologies, we develop methods for

• Logic synthesis under technology constraints,
• Placement & routing for technology-independent layout generation,
• Clocking and data synchronization,
• Layout validation and verification, and
• Technology mapping and optimization

For more details on our work in the domain of field-coupled nanotechnologies, please see this page.

Besides that, we also successfully started according research for further alternative technologies, namely

• the design of adiabatic circuits,
• the design of optical circuits, and
• the design of memristor circuits.

Contributions to Complementary Research Areas

Besides the research areas sketched above, we also try to apply our expertise in complementary areas. This led to further contributions to the fields of

• High Performance Computing for fluidic simulation,
• Machine Learning for hardware design,
• Algorithms for automatic formal proof techniques,
• Studies on graph transformations,
• Formalization of legal regulations, and
• others.