主办：陕西省汽车工程学会
ISSN 1671-7988  CN 61-1394/TH
创刊：1976年
汽车实用技术

汽车实用技术 ›› 2022, Vol. 47 ›› Issue (23): 45-50.DOI: 10.16638/j.cnki.1671-7988.2022.023.009

• 智能网联汽车 • 上一篇    

基于视觉的车道保持系统

张杭铖，杜海林，王齐超   

  1. 长安大学 汽车学院
  • 出版日期:2022-12-15 发布日期:2022-12-15
  • 通讯作者: 张杭铖
  • 作者简介:张杭铖（1998—），男，硕士研究生，研究方向为无人驾驶，E-mail:1135498030@qq.com。

Vision-based Lane Keeping System

ZHANG Hangcheng, DU Hailin, WANG Qichao   

  1. College of Automobile, Chang’an University
  • Online:2022-12-15 Published:2022-12-15
  • Contact: ZHANG Hangcheng

PDF

摘要： 为提高无人驾驶车辆主动安全性，文章基于神经网络车道线检测和单点预瞄横向控制 策略的车道保持方法，以车载摄像头作为感知设备，采集前方道路的图像信息并传输给车载 终端对图片进行处理后得到车道线信息，据此设计相应的横向控制器计算出偏离车道中心所 需的前轮转角，最终作为控制信号传输给车辆的控制机构实现车道保持。经过实车验证表明， 在低速情况下该系统能实现车道保持功能。研究结果对提高汽车的车道保持性能有一定参考 价值。

关键词: 神经网络；车道线识别；横向控制；车道保持

Abstract: In order to improve the active safety of unmanned vehicles, this paper adopts the lane keeping method based on neural network lane line detection and single point preview lateral control strategy. The on-board camera is used as the sensing device to collect the image information of the road ahead and transmit it to the on-board terminal to process the image and obtain the lane line information. Accordingly, the corresponding lateral controller is designed to calculate the front wheel Angle required to deviate from the lane center, which is finally transmitted to the vehicle control mechanism as a control signal to realize lane keeping. The test results show that the system can realize the lane keeping function at low speed. The research results have some reference value for improving the performance of lane keeping.

Key words: Neural network; Lane line recognition; Transverse control; Lane keep