中文核心期刊
CSCD来源期刊
中国科技核心期刊
RCCSE中国核心学术期刊

重庆交通大学学报(自然科学版) ›› 2021, Vol. 40 ›› Issue (03): 27-33.DOI: 10.3969/j.issn.1674-0696.2021.03.05

• 交通+大数据人工智能 • 上一篇    下一篇

基于自学习的组网式交通信号灯异常检测研究

刘永涛1,2,樊亚敏1,张莉1,黎冠1   

  1. (1. 华北科技学院 电子信息工程学院,河北 燕郊 065201; 2. 中国农业大学 信息与电气工程学院,北京 100083)
  • 收稿日期:2019-06-20 修回日期:2019-10-16 出版日期:2021-03-15 发布日期:2021-03-15
  • 作者简介:刘永涛(1981—),男,河北定州人,副教授,博士,主要从事物联网和机器人技术方面的研究。E-mail:ytliu@ncist.edu.cn 通信作者:黎冠(1981—),男,山东成武人,副教授,博士,主要从事智能机器人技术方面的研究。E-mail:123837664@qq.com
  • 基金资助:
    中央高校基本科研业务费资助项目(3142018047);河北省重点研发计划自筹项目(17273908)

Anomaly Detection of Traffic Signal Lights in Network Based on Self-learning

LIU Yongtao1, 2, FAN Yamin1, ZHANG Li1, LI Guan1   

  1. (1. School of Electronic Information Engineering, North China Institute of Science and Technology, Yanjiao 065201, Hebei, China; 2. School of Information and Electrical Engineering, China Agricultural University, Beijing 100083, China)
  • Received:2019-06-20 Revised:2019-10-16 Online:2021-03-15 Published:2021-03-15
  • Supported by:
     

摘要: 随着智慧交通的发展,信号灯成为城市道路能够正常运转的关键性设备,交通部门有必要实时掌握信号灯的运行状态,以便及时发现故障并加以排除,降低交通事故的发生率。针对现有研究方法存在的硬件结构复杂、成本较高、安装施工困难、维护不便等问题,提出了一种基于自学习的组网式信号灯故障诊断方法。系统可以一键学习、快速准确地掌握正常情况下交通信号灯的运行状态,通过动态阈值模糊推理诊断方法,解决了由于系统老化、温度变化等原因造成的阈值漂移问题。整体采取无线组网的方式和一拖三测量结构,通过边缘计算分布式处理的方法,很大程度上降低了施工难度,硬件系统成本降低50%以上,有效解决了当前设计、安装存在的诸多问题。

 

关键词: 交通运输工程, 智慧交通, 自学习, 无线组网, 边缘计算, 动态阈值, 希尔排序

Abstract: With the development of intelligent traffic, signal lights have become the key equipment for the normal operation of urban roads. It is necessary for traffic departments to grasp the operation status of signal lights in real time to find out and eliminate the faults in time and reduce the incidence of traffic accidents. Aiming at the problems of complex hardware structure, high cost, difficult installation and construction, and inconvenient maintenance of existing research methods, a fault diagnosis method of networked signal lights based on self-learning was proposed. The system could quickly and accurately grasp the operation status of signal lights under normal conditions by one click learning. By means of dynamic threshold fuzzy reasoning and diagnosis method, the problem of threshold drift caused by system aging and temperature change was solved. The whole system adopted wireless networking mode and one-drag-three measurement structure. Through the edge computing distributed processing method, the construction difficulty was greatly reduced and the cost of hardware system was reduced by more than 50%, which effectively solved many problems existing in current design and installation.

Key words: traffic and transportation engineering, intelligent traffic, self-learning, wireless networking, edge computing, dynamic threshold, Shell sorting

中图分类号: