Simulation of Mixed Critical In-vehicular Networks

TL;DR

This paper discusses the simulation of complex in-vehicular networks using domain-specific models, focusing on Ethernet-based communication with real-time QoS, to support future automotive applications like autonomous driving.

Contribution

It introduces domain-specific simulation models and workflows for designing in-vehicular networks with Ethernet and TSN extensions, demonstrated through a case study on a premium car.

Findings

Developed simulation models for in-vehicular Ethernet networks.

Applied models to a real-world car network case study.

Enhanced understanding of network design for future automotive systems.

Abstract

Future automotive applications ranging from advanced driver assistance to autonomous driving will largely increase demands on in-vehicular networks. Data flows of high bandwidth or low latency requirements, but in particular many additional communication relations will introduce a new level of complexity to the in-car communication system. It is expected that future communication backbones which interconnect sensors and actuators with ECU in cars will be built on Ethernet technologies. However, signalling from different application domains demands for network services of tailored attributes, including real-time transmission protocols as defined in the TSN Ethernet extensions. These QoS constraints will increase network complexity even further. Event-based simulation is a key technology to master the challenges of an in-car network design. This chapter introduces the domain-specific aspects and simulation models for in-vehicular networks and presents an overview of the car-centric network design process. Starting from a domain specific description language, we cover the corresponding simulation models with their workflows and apply our approach to a related case study for an in-car network of a premium car.