Search Results

As demand for ADAS and autonomous systems grows, as well as the need to avoid lawsuits from software bugs, the desire for high integrity and functional safety applications is growing. While there are many process and technical factors that contribute to the overall quality of the software, it is becoming increasingly important to define the software architecture and verify that the implementation satisfies all of the architecture requirements. Because defining the software architecture is new for many companies, there is a need to know what properties are important to specify. Also lacking are techniques needed to verify that the implementation is a “good” architecture.

This work presents current methods to analyze and improve the architecture of Simulink models. The methods follow the “principles for architectural design” of part 6 on software development of the ISO 26262 functional safety standard for road vehicles, the dominating standard in the automotive industry. The methods presented describe how the abstract architectural principles of the ISO 26262 can be implemented in the context of model-based development using Simulink. Therefore we demonstrate how different metrics can be used to improve or enforce the compliance with the principles. In contrast to previous publications we will not primarily focus on the metrics itself, but emphasize the architectural principles themselves and expose the architectural implications of applying the metrics. As the architectural principles of the ISO 26262 are targeted at reducing the overall complexity, we will also focus on metrics and methods that help to reduce the models complexity.