Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Adaptive QoS management in automotive in-vehicle networks

A new approach enabling the efficient planning of network resources for packet based in-vehicle networks
: Manderscheid, Martin; Plankl, Arnold; Langer, Falk; Engelhardt, Daniel

Postprint urn:nbn:de:0011-n-1866405 (110 KByte PDF)
MD5 Fingerprint: 3f149517ea802048e767e75d3822bb93
Erstellt am: 8.12.2011

VDI/VDE-Gesellschaft Meß- und Automatisierungstechnik -GMA-, Düsseldorf; VDI-Wissensforum GmbH:
Autoreg 2011 - Steuerung und Regelung von Fahrzeugen und Motoren : 5. Fachtagung Baden-Baden, 22. und 23. November 2011
Düsseldorf: VDI-Verlag, 2011 (VDI-Berichte 2135)
ISBN: 978-3-18-092135-8
ISSN: 0083-5560
14 S.
Fachtagung AUTOREG <5, 2011, Baden Baden>
Konferenzbeitrag, Elektronische Publikation
Fraunhofer ESK ()
quality of service; in-vehicle communication; network states; ethernet/IP

Analyzing the use of Ethernet and the Internet Protocol (IP) inside the car, network planning and validation turned out to be one of major tasks to be solved. While several works showed that Ethernet can be used in general inside a car, questions concerning a suitable design process are still remaining open. In previous works two approaches have been proposed for the design process of automotive in-vehicle networks. The first method exploits the fact that many automotive functions do not communicate at the same time. The modeling of the time dependencies using binary relations has been proposed. On this foundation a formal way is provided to detect which functions will communicate simultaneously. The second method shows a way to validate the communication requirements of Ethernet based in-vehicle network configurations using the Network Calculus. Combining those two approaches a formal process validating Ethernet network configurations is developed. This process will be presented in this work. Furthermore, an exemplary case study will be presented, emphasizing the necessity of a formal process. Validating the case study's network configuration, the end to end delays of the functions are calculated in two different ways. For the first way the before reviewed approach is applied. I.e. the time related behaviour of the functions is examined, thus allowing differentiating which functions will communicate at the same time. For the second way a straight forward approach is used. Comparing those results, it becomes clear which effects can occur while ignoring (willingly or not) the time dependencies between electronic functions.