The number of functions in the vehicle has been growing constantly for the last 50 years and is expected to grow further, driven by innovation, competition and stricter regulations. Also the number of electronic control units (ECUs) has grown to reach 70 ECUs in the high-end segment. A further increase would have negative effects on the vehicle cost, weight and fuel consumption, so that the integration of functions in less ECUs has become necessary. while enabling technologies such as the AUTomotive Open System ARchitecture (AUTOSAR) software platform and powerful hardware platforms are available, the design engineers face a high number of integration alternatives and competing design objectives making architectural decisions very complex and challenging. ln order to meet the complexity of the architecture design task, a new approach to the automated design and optimisation of network architectures will be presented. The requirements for the network architecture are classified in terms of what must be fulfiled fully (design constraints) and what can be optimised (design objectives). The fulfilment degree of design objectives (e.g., cost efficiency, extensibility) varies from a design alternative to another, so that metrics are required to measure this variation. While established metrics are available for reliability and cost efficiency, new metrics will be introduced for the efficiency of variants management and extensibility. The design of the network architecture can be considered as a combination of two clustering problems: the clustering of application components into HW units (e.g. ECUs) and the clustering of the so obtained HW units and additional standalone gateways into sub-networks. The combined space of possible design alternatives with respect to both clustering problems is very large and complex. Therefore, a two-level multi-objective evolutionary optimisation algorithm is suggested in order to assist the design engineers in exploring this design space with respect to the chosen design objectives and under the specified constraints. Since multiple design objectives are considered, a set of competing design alternatives results from optimisation. These design alternatives correspond different trade-offs between design objectives. Therefore a decision method is developed in order to assist the design engineers in the final architectural design decision by providing them with an overall ranking of the generated alternatives. This method is based on expert knowledge from design engineers and on analytical methods. ln order to validate and verify the concepts presented in this work, a case study is carried out. This study shows that the new method allows for the automated generation and comparison of different network topologies. It also delivers a set of optimal design alternatives that can compete with the state-of-the-art designs and even represent a realistic alternative to them with regard to the considered design objectives.


    Access

    Access via TIB

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    A new approach for gateway-based automotive network architectures


    Contributors:


    Publication date :

    2009


    Size :

    154 Seiten, Bilder, Tabellen, Quellen



    Type of media :

    Theses


    Type of material :

    Print


    Language :

    English






    FULL FEATURED PACKET-BASED AUTOMOTIVE NETWORK SECURITY GATEWAY

    QIAO MING / SMITH JAN | European Patent Office | 2020

    Free access

    Gateway Application for Automotive Network System "BEAN"

    Tanaka, A. / Kamiya, M. / Sakai, K. et al. | British Library Conference Proceedings | 1999


    Gateway Application for Automotive Network System “BEAN”

    Tanaka, Akihiro / Sakai, Kazunori / Kamiya, Masachika et al. | SAE Technical Papers | 1999