Design of Robust System Architectures for Automotive ECUs

The Paper describes a model based approach to improve the robustness and reliability of an embedded system with respect to real-time performance. It characterizes how real-time simulation models are generated, run and analyzed to gain knowledge of the dynamic system behavior. The system’s reactions to dynamic stimuli can be predicted without having to implement all hardware and software. A study with an automotive car body control unit illustrates how the timing model is developed parallel to the development progress. Findings and improvements are listed. Besides the technical aspects, the business impact for the current and future systems proves significant advantages of the chosen approach.