汽车工程 ›› 2023, Vol. 45 ›› Issue (12): 2338-2347.doi: 10.19562/j.chinasae.qcgc.2023.12.016

所属专题: 智能网联汽车技术专题-控制2023年

• • 上一篇    下一篇

基于改进连续型Hopfield神经网络的CAN总线负载率优化

何智成1,杜磊浩2,周恩临1(),覃高峰2,黄晋3   

  1. 1.湖南大学,汽车车身先进设计制造国家重点实验室,长沙 410082
    2.上汽通用五菱汽车股份有限公司,柳州 545007
    3.清华大学车辆与运载学院,北京 100084
  • 收稿日期:2022-04-19 修回日期:2022-06-12 出版日期:2023-12-25 发布日期:2023-12-21
  • 通讯作者: 周恩临 E-mail:tenrey18@163.com
  • 基金资助:
    国家自然科学基金联合基金(U20A20285)和湖南省杰出青年基金(2021JJ10016)资助。

CAN Bus Load Rate Optimization Based on Improved Continuous Hopfield Neural Network

Zhicheng He1,Leihao Du2,Enlin Zhou1(),Gaofeng Qin2,Jin Huang3   

  1. 1.Hunan University,State Key Laboratory of Advanced Design and Manufacturing for Vehicle Body,Changsha  410082
    2.SAIC GM Wuling Automobile Co. ,Ltd. ,Liuzhou  545007
    3.School of Vehicle and Mobility,Tsinghua University,Beijing  100084
  • Received:2022-04-19 Revised:2022-06-12 Online:2023-12-25 Published:2023-12-21
  • Contact: Enlin Zhou E-mail:tenrey18@163.com

摘要:

CAN总线负载率对于车载总线的安全性与时延性有至关重要的作用。鉴于传统的连续型Hopfield神经网络(CHNN)在解决此类问题时存在惩罚参数鲁棒性差和所得解易陷入局部最优的缺陷,本文借助模拟退火算法中的蒙特卡洛思想,提出应用于CAN总线负载率优化问题的改进连续性Hopfield神经网络算法(SA-CHNN)。选取微型电动车中99条通信信号作为实验数据进行测试,结果表明,SA-CHNN算法成功解决传统CHNN算法求解CAN总线负载率优化问题的不足,具有明显的优越性。最后,基于搭建的Simulink-Speedgoat CAN总线实验平台对SA-CHNN算法得到的最优信号分配的报文进行实时负载率仿真验证,结果表明SA-CHNN算法的准确性。

关键词: CAN总线, 负载率, 神经网络, 模拟退火法, Metropolis准则

Abstract:

The Busload rate of CAN bus is of vital importance to the security and delay bounds of Vehicle-Bus. Whereas, the traditional continuous Hopfield neural network (Continuous Hopfield Neural Network, CHNN) has the defects of poor penalty parameter robustness and the resulting solution that easy to be trapped in local optimality when solving such problems, based on the Metropolis thought in the simulated annealing algorithm, an improved continuous Hopfield neural network algorithm (SA-CHNN) applied to the CAN bus load rate optimization problem is proposed in this paper. Micro-electric vehicle’s ninety-nine communication signals are selected and tested as experimental data. The results show that the SA-CHNN algorithm successfully solves the problems of the traditional CHNN algorithm in solving the CAN bus load rate optimization, which has obvious advantages. Finally, based on the Simulink-Speedgoat CAN bus experimental platform, the real-time load rate simulation with the optimal signal distribution messages is conducted and the result reveals the accuracy of the SA-CHNN algorithm.

Key words: CAN bus, load rate, neural network, simulate annealing algorithm, Metropolis guidelines