Master Theme Embedded Systems

Coordinator: prof.dr. Frits Vaandrager

Embedded systems is one of the possible specialisation themes in the Computer Science Master Programme at the Radboud University Nijmegen. It relates closely to the research carried out within the Model-Based System Development (MBSD) group. What are Embedded Systems? Embedded systems are highly specializable, often reactive, sub systems that provide, unnoticed by the user, information processing and control tasks to their embedding system. Embedded systems are omnipresent nowadays and make possible the creation of systems with a functionality that cannot be provided by human beings. Example application areas are consumer electronic products (eg CD players, microwave ovens), telecommunication (eg mobile phones), medical systems (eg pacemakers), traffic control (eg intelligent traffic lights), driving and car control (eg ABS), airborne equipment (eg fly-by-wire), and plant control (eg packaging machines, wafer steppers). The term embedded system thus encompasses a broad class of systems, ranging from simple microcontrollers to large and complex multi-processor and distributed systems.

A Challenging Task The importance of embedded systems is undisputed. Their market size is about 100 times the desktop market, and will grow exponentially also in the next decade. Ever more intelligence will be added to processes and products, and eventually they will all be able to communicate with each other via the Internet. Characteristics of embedded systems are: Complex interaction with environment. One can only design and reason about embedded systems if one takes the behavior of their environment into account. Frequently this environment is highly nondeterministic and intrinsically continuous. High dependability requirements. Besides functional constraints many other aspects play a role in the design of embedded systems: timeliness, fault tolerance, availability, security, safety, etc. This combination of factors makes the design of embedded systems in general a very complex task. Failure of embedded systems often may have serious consequences (loss of lives, huge financial losses), so correctness and reliability are of vital importance.

The MBSD group develops methods and tools for the specification, design, analysis and testing of embedded systems, distributed algorithms and protocols, and assesses and demonstrates the effectiveness of these methods and tools in the industrial development process. MBSD has an excellent international reputation and its research program was singled out as "Excellent" in the last national research assessment. Members of the MBSD group closely collaborate with industry in joint research projects (Philips, OCE, ASML, Siemens, Chess, Bosch, and with the Embedded Systems Institute (ESI)).

Courses

• NWI-I00155 Design of Embedded Systems (DES) (Jozef Hooman)
  General aim of this course is to acquire basic knowledge about the systematic design of embedded systems and to obtain practical experience with the development of such systems. Relevant topics are, for instance, timing, real-time operating systems (and experience with Real-Time Linux), scheduling, interface of software with sensors and actuators, and the development of a small distributed real-time application. As a case study you build a (Lego) Mars Rover (in our Robot LAB) which has to accomplish a certain mission. Click here to see some movies demonstrating the work of this year's students.
• NWI-I00154 Analysis of Embedded Systems (AES) (Frits Vaandrager)
  During recent years there has been enormous progress in the areas of hardware and software verification and analysis. In this course an overview will be presented of mathematical techniques for the specification and analysis of embedded systems. The application of these techniques will be illustrated on industrial sized examples taken from the areas of embedded real-time systems, distributed algorithms and protocols. Participants learn how to make formal models, and how to analyze them using state-of-the-art techniques and tools. In 2006, for instance, the students modeled a part of the Bluetooth protocol using the tool Uppaal. They wrote a paper on this, which was presented at a conference in Canada. In 2009, students wrote 2 papers on wireless sensor networks and scheduling of printers that were presented at an internation workshop. For popular accounts of the research area covered by this course see an interview in the November issue of CHIP, a recent paper that Frits Vaandrager wrote for Bits & Chips, an article in NRC Handelsblad (pdf), or listen to an interview on VPRO radio (all in Dutch). In the course evaluation by students, the Analysis of Embedded Systems Course consistently gets very high overall grades (8.8 in 2009).
• NWI-I00110 Testing Techniques (Jan Tretmans)
  The course "testing techniques" deals with a number of techniques, methods and tools which may help in the systematic and effective testing of software systems. Established testing techniques as well as some new developments, such as testing with formal specifications and model-based testing, will be presented. Some guest lectures will be given by people from industrial software testing.
• NWI-IMC037 Computer aided diagnosis in medical imaging (Nico Karssemeijer)
  Imaging is increasingly important in modern medicine. In screening, diagnosis, and treatment, innovative procedures based on imaging are gradually changing the way healthcare is practiced. As the workload of clinicians and the complexity of imaging increases, there is a growing need for more intelligent processing, to aid clinicians with image interpretation and to provide decision support tools. A first generation of such systems has been introduced successfully in practice, in applications ranging from advanced visualization to computer aided detection and diagnosis (CAD). In particular in breast cancer screening automated detection of abnormalities is already used on a large scale. To further develop this technology, the development of knowledge based systems, in which image analysis is combined with knowledge extracted from large imaging archives and associated medical records, is crucial.