Methodik zur ganzheitlichen, dynamischen Analyse sicherheitsrelevanter, verteilter Kraftfahrzeugsysteme unter dem Aspekt des Fehlerverhaltens (English shop)

Preview

Table of Contents, Datei (43 KB)
Preface, Datei (29 KB)
Extract, Datei (75 KB)

In recent years, a clear trend towards more electronics, highly networked structures and complex assistance and safety systems has become apparent in the automotive industry.

These systems support the driver in critical situations and relieve them of the responsible task of operating a motor vehicle. The increasingly extensive transfer of responsibility from the human being to technical systems places high demands on system safety in order to limit the risk of a hazard originating from the motor vehicle system. With the aid of safety analyses, the risk present in the operation of motor vehicle systems can be analysed. Classical safety analyses generally consider only individual constituents of a motor vehicle system, which, however, in general comprises the following main constituents: the motor vehicle system itself, the vehicle, the vehicle environment and the driver. But what happens if a component in the motor vehicle system fails, the software contains faults and at the same time the driver initiates a critical driving manoeuvre? Such questions can only be answered inadequately by classical safety analyses. It should furthermore be noted that in most classical safety analyses the actual analysis is carried out in the form of brainstorming processes with teams of experts. However, no expert is able to survey and assess all possible fault effects and fault combinations in the interplay of the complex constituents. The present work presents a model-based approach for conducting a holistic safety analysis which takes all constituents of a motor vehicle system into account. This approach extends the methods of model-based systems development used in the automotive industry to safety aspects and combines them with classical safety analyses. The modelling of the overall system is based on a quantitative, dynamic and continuous approach such as is also employed in systems development for rapid prototyping purposes. The presented analysis is carried out computer-aided and, with the aid of an identification algorithm, can autonomously detect, already during system development, single faults, fault combinations and fault sequences that lead to significant effects in the overall system. Through iterative application of this approach it is possible to analyse the motor vehicle system continuously, to identify potential for improvement at an early stage and to implement corresponding changes in a targeted manner.