基于多功能车辆总线事件仲裁的优化设计

Optimized Pattern Based on MVB Event Arbitration

事件轮询机制是多功能车辆总线MVB实现消息数据通信的关键,为实时协议RTP中消息服务的高效网络通信提供链路层的支持。事件仲裁算法是事件轮询机制的核心构成,在研究事件轮询流程机理的基础上,根据IEC61375-1标准规定的基本仲裁算法,引入仲裁二叉树模型,实现事件仲裁算法并成功应用于列车控制网络平台中。在仲裁二叉树模型下,IEC61375-1推荐的连续地址分配方式不能保证事件仲裁时延的最优性,提出在有限深度的条件下运用哈夫曼算法生成最优二叉树分配设备地址。通过对大量事件突发和车辆设备数量分布不均情况的仿真,比较两种分配方式的时延性能,分析有限深度最优二叉树算法的具体优势。

Event polling is pivotal in MVB message communication. It supports the efficiency of link layer in RTP specified message service. Event arbitration is the core of event polling. Analyzing the event polling mechanism and basing on the general arbitration algorithm specified in IEC61375-1 ,the arbitration binary tree model was introduced and the event arbitration algorithm was accomplished and applied on the vehicle control network platform. With the arbitration binary tree model, the IEC61375-1 recommended consecutive-address allocation pattern cannot guarantee the optimal time delay of event arbitration. So the Huffman algorithm was proposed to be used under the condition of limited depths to generate the optimal binary tree device allocation addresses. By simulating event explosion and imbalanced device distribution, the time delay performance of the above two patterns were compared and the specific superiority of the new pattern was detailed.