Software Engineering for Embedded Systems (WS 2011/12)

Prof. Dr. Holger Giese


Software-intensive systems, in which a considerable fraction of the system development efforts is dedicated to the development of embedded software, are often regarded as the most important software engineering field in the years to come. They are expected to be one key factor of success for many industries such as, for example, the automotive sector, transportation, or medical devices. As today technical systems also become connected to each other using network technology, we no longer only have technical systems which are controlled by isolated operating embedded software. Instead, the software may include complex information processing capabilities and the coordination between the different technical systems via networks taking hard real-time constraints into account.

Modeling embedded systems often results in a mix of models from a multitude of disciplines such as software engineering, control engineering, mechanical engineering, and electrical engineering. Block diagrams in systems engineering and the Unified Modeling Language (UML) in software engineering, are prominent examples of domain specific modeling techniques used for modeling. Recently, several steps towards integrating both worlds can be observed. UML 2.0 component diagrams offer a system view which has been originally invented for complex real-time systems in the telecommunications domain. SysML suggests an extension of the UML for systems engineering. The required integration has to combine the usually continuous world considered by systems engineering and the discrete software engineering view and thus results in techniques for hybrid systems which support both continuous as well as discrete behavior.

Embedded systems are often safety-critical applications where their correct operation is vital to ensure the safety of the public and environment. Examples include shut-down systems for nuclear power plants, fly-by-wire aircrafts, autonomous train control software or anti-lock braking systems in automobiles. Safety is a system property and thus cannot be studied by simply taking into account the software part of an embedded system, only. However, in this lecture we will address the general engineering aspects of safety in a rather superficial manner and mainly concentrate on the specific problems of safety-critical systems which contain (complex) software parts.

In this lecture, we will review the current state of the art of software engineering for embedded systems taking into account the techniques available for the different development activities such as project management, requirements engineering, analysis & design, implementation, and verification & validation. This will, in particular, include the study of available techniques for the development of systems, which are safety-critical, have hard real-time constraints, and are hybrid systems. Also an overview about the current state of the art for the model-driven development of embedded systems is provided. In addition to the lecture exercises are organized to give an insight how to use state of the art approaches and tools. Within small projects the students can contribute the gained knowledge by developing solutions for the Robotino-Robot by using these introduced tools and concepts. The exercises and projects will be organized (partially) in the context of the currently established new Cyber-Physical-Systems-Laboratory (CPS-Lab) at the Hasso-Plattner-Institute.

Introduction

Foundations

Application Domains & Life Cycle

Exercises

Requirements

Date: November 9, 2011
Language: English
Duration: 01:27:48

Analysis & Design

Date: November 22, 2011
Language: English
Duration: 00:52:36

Implementation