Driver assistance features are increasingly dependent upon system architectures that distribute and share responsibilities across various function-based ECUs to minimize cost and redundancy while maximizing engineering efficiency. Clear and accurate system requirements are critical to success, and a robust methodology for validating and testing requirements is essential. Distributed systems are highly sensitive to requirement ambiguity and inaccuracy as they are designed on the assumptions of predictable logical behavior of each functional component. Requirement ambiguity drives variance in implementations which results in system incompatibilities. Errors in requirements lead to faulty implementations that fail not just the component test but also hinder the testing of the entire system of components.
This paper shall discuss a Model-Based Design & Systems Engineering methodology which emphasizes modeling system requirements up-front in a unified simulation environment in order to conduct executable tests of the requirements and validate they deliver the design intent. It shall provide an overview of the requirement modeling methodology, and give real application examples from the development of the 2016MY Ford F-150 Pro Trailer Backup Assist™ feature to demonstrate how requirement modeling can be applied to discover and eliminate requirement and implementation defects through virtual testing and model-based design.
