汽车工程

汽车工程 ›› 2023, Vol. 45 ›› Issue (5): 746-758.doi: 10.19562/j.chinasae.qcgc.2023.ep.007

所属专题： 智能网联汽车技术专题-控制2023年

• • 上一篇    下一篇

面向时间敏感网络的车载以太网网络架构多目标优化

邹渊1,2,孙文景1,2,张旭东1,2(),温雅1,2,曹万科1,2,张兆龙3   

  1. 1.北京理工大学机械与车辆学院，北京 　100081
    2.北京理工大学，电动车辆国家工程研究中心，北京 　100081
    3.北京新能源汽车股份有限公司，北京 　100176
  • 收稿日期:2022-10-26 修回日期:2022-11-29 出版日期:2023-05-25 发布日期:2023-05-26
  • 通讯作者: 张旭东 E-mail:xudong.zhang@bit.edu.cn
  • 基金资助:
    国家重点研发计划(2021YFB2500900)

Multi-objective Optimization of In-Vehicle Ethernet Network Architecture for Time-Sensitive Network

Yuan Zou1,2,Wenjing Sun1,2,Xudong Zhang1,2(),Ya Wen1,2,Wanke Cao1,2,Zhaolong Zhang3   

  1. 1.School of Mechanical Engineering，Beijing Institute of Technology，Beijing 　100081
    2.Beijing Institute of Technology，National Engineering Research Center for Electric Vehicles，Beijing 　100081
    3.Beijing Electric Vehicle Co. ，Ltd. ，Beijing 　100176
  • Received:2022-10-26 Revised:2022-11-29 Online:2023-05-25 Published:2023-05-26
  • Contact: Xudong Zhang E-mail:xudong.zhang@bit.edu.cn

RichHTML

11

PDF (PC)

200

赞

摘要：

车载电子电气架构的网络架构深刻影响智能网联车辆的通信安全性和确定性。针对面向时间敏感网络（TSN）的区域-功能域电子电气架构，本文首次建立了以端口数均匀、负载均衡和信息流端到端延时最低为优化目标的网络架构多目标优化框架。通过求解TSN流量调度问题获得信息流延时，将流量调度抽象为周期性车间作业调度问题（JSP），提出适用于流量调度的多种群遗传算法（MPGA），相比于传统遗传算法求解效果提高16%。为了快速求解多目标优化问题，设计了改进的快速非支配排序遗传算法（NSGA-II），通过引入迭代因子和拥挤因子对算法进行自适应交叉变异概率改进，优化效率提高了25%。仿真验证了多目标优化框架的有效性并为面向TSN的车载以太网网络架构优化提供了一种设计思路。

关键词: 电子电气架构, 网络架构, 时间敏感网络, 流量调度, 多目标优化

Abstract:

The network architecture of the vehicle electrical and electronic architecture profoundly affects communication security and certainty. For Zone-Domain based electrical and electronic architectures that use time-sensitive networks （TSN）， this paper establishes for the first time the multi-objective optimization framework for network architecture with the optimization objectives of uniform number of ports， balanced load and lowest end-to-end delay of information flows. The end-to-end delay is obtained by solving the TSN traffic scheduling and the traffic scheduling is abstracted as the periodic job-scheduling problem （JSP）. The multi-population genetic algorithm （MPGA） applicable to traffic scheduling is proposed， which improves the solution effect by 16% compared with the traditional genetic algorithm. In order to solve the multi-objective optimization problems rapidly， an improved non-dominated sorting genetic algorithm （NSGA-II） is designed in this paper. The optimization efficiency is improved by 25% by introducing in the iteration factor and congestion factor to improve the algorithm with adaptive cross-variance probability. The simulation verifies the effectiveness of the multi-objective optimization framework and provides a design idea for the optimization of in-vehicle Ethernet network architecture with the introduction of TSN.

Key words: electrical and electronic architecture, in-vehicle network architecture, time-sensitive network, traffic scheduling, multi-objective optimization