Scholarlink

This paper proposes a low-cost intelligent vehicle control system based on a Raspberry Pi 4B embedded platform (Broadcom BCM2711 quad-core, VideoCore VI GPU), which employs a lightweight image processing technique to detect lane lines in real time and perform path tracking. The system combines the ARM architecture optimized PD control algorithm with OpenCV4-Tegra to achieve a decision delay of 200ms at a low power consumption of 5W, with an accuracy of 92% for crosswalk detection and automatic stopping function. Experimental results show that the path tracking deviation is stabilized within 5%, the cost of the scheme is reduced by 67% compared to Jetson, and the memory occupation is reduced by 40%.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Pi 4B-Based Visual Navigation System for Low-Computing Power Intelligent Vehicles

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Jiahui Yang,
  • Shengtong Mai,
  • Jinbo Cai,
  • Jiaxian Chen,
  • Jinwen Yang

摘要

This paper proposes a low-cost intelligent vehicle control system based on a Raspberry Pi 4B embedded platform (Broadcom BCM2711 quad-core, VideoCore VI GPU), which employs a lightweight image processing technique to detect lane lines in real time and perform path tracking. The system combines the ARM architecture optimized PD control algorithm with OpenCV4-Tegra to achieve a decision delay of 200ms at a low power consumption of 5W, with an accuracy of 92% for crosswalk detection and automatic stopping function. Experimental results show that the path tracking deviation is stabilized within 5%, the cost of the scheme is reduced by 67% compared to Jetson, and the memory occupation is reduced by 40%.