智能汽车Ethernet AVB网络实时性分析

Real-time Analysis of Ethernet AVB Network for Intelligent Vehicle

针对Ethernet AVB网络的实时性进行研究,通过分析某智能汽车的Ethernet AVB网络拓扑结构,定义消息参数模型,同时根据实时性要求将消息进行分类,并提出预定流量模型以传输时间关键的控制类消息。在此基础上,通过分析Ethernet AVB网络协议提出每类消息的最坏响应时间概念,并推导出定量计算公式。针对该最坏响应时间计算方法进行优化,以获得更加严格的最坏响应时间上界限。在Vector CANoe中的仿真结果表明,该方法不仅可以保证最坏响应时间的严格上界限,而且还可以有效减少上界限平均约33.7%。

The real-time of Ethernet AVB network is studied in this paper. Based on the analysis of the Ethernet AVB network topology for an intelligent vehicle, the network message parameter model is defined. And the messages are classified according to real-time requirements, then a scheduled traffic model is proposed to transmit time-critical control message. On this basis, the concept of the worst response time for each type of message is proposed by analyzing the Ethernet AVB network protocol, and the quantitative calculation formula is deduced. The worst-case response time computation method is optimized to obtain a tighter upper bound of the worst-case response time. The simulation results in Vector CANoe show that this method can not only guarantee the tight upper bound of the worstcase response time, but also effectively reduce the upper bound about 33.7% on average.