Improving the reliability and performance of FlexRay vehicle network applications using simulation techniques

Abstract Modern vehicles are becoming more and more sophisticated, with more functions being controlled by a microprocessor unit. As new functions are developed there is not only more of a demand on the control unit, but there is also more demand placed on the communication network(s) within a car. There is also a growing need for fast and dependable networks for new safety features such as X-by-wire applications. A trend in the automotive industry to make cars more eco-friendly has emerged. As the amount of applications increases then the number of wires within a car increases and this can potentially add a large amount of weight leading to increased fuel consumption. This along with the need for higher performance networks led to the development of the FlexRay protocol. FlexRay is a newly developed network protocol that is intended to address the current and future needs of the automotive industry. It is backed by many automotive manufacturers and suppliers. As such, FlexRay looks increasingly likely to become the network application of choice for many companies, especially where safety critical systems are implemented. The purpose of this research was to design, implement and test a simulation model of a FlexRay network node. This simulation model could be a benefit to system developers to ensure accurate communication is achieved by tracing the flow of information through a FlexRay-based system and ensuring all timing constraints are met. The model was built using MATLAB, Simulink and SimEvents. The basis of the model was a node that incorporated a separate host microcontroller and communications controller. The communications controller was based on The Bosch E-Ray IP. The simulation model comprised of the application, software driver, communications and physical bus layers of a FlexRay based system. The model was then calibrated against a real world system over a number of different test cases and constraints. The final part of the research involved running tests to determine if the model that has been built, was built in a correct manner i.e. validation of the simulation model. The model was then evaluated for its ability to carry out its intended role.