Communications-Oriented Modeling and Development of Vehicular Distributed Embedded Systems

Sammanfattning: The model- and component-based development approach has emerged as an attractive option for the development of vehicular distributed real-time embedded systems. Within this context we target challenges related to modeling of legacy network communication, extraction of end-to-end timing models and support for end-to-end timing analysis.We propose a novel approach for modeling legacy network communication in these systems. By introducing special-purpose components to encapsulate and abstract the communication protocols, we allow the use of legacy nodes and legacy protocols in a component- and model-based software engineering environment. Because an end-to-end timing model should be available to perform the end-to-end response-time and delay analyses, we present a method to extract the timing models from these systems. We also extend the method to various abstraction levels and parts of the development process for the systems. During the models extraction, we identify that the existing worst-case response-time analysis for Controller Area Network (CAN), a widely used real-time network protocol in the vehicular domain, does not support mixed messages. These messages are partly periodic and partly sporadic. They are implemented by some higher-level protocols for CAN used in the industry. We extend the existing analysis which is now applicable to any higher-level protocol for CAN that uses periodic, sporadic and/or mixed transmission.In order to show the application of our modeling techniques, timing model extraction method and extended analyses; we provide a proof of concept by extending the Rubus Component Model, which is used for the development of software for vehicular embedded real-time systems by several international companies. We also implement the end-to-end response-time and delay analyses along with the extended analysis for CAN in the existing industrial tool suite the Rubus-ICE. Moreover, we implement the extended analysis for CAN in a free tool MPS-CAN analyzer. Further, we conduct automotive-application case studies to validate our methods and techniques.