Scholarlink

Traffic congestion in urban areas is an increasing challenge because of fixed traffic signals, which do not change according to real conditions. This paper brings in an Adaptive Traffic Light Control System based on computer intelligence and vision, allowing for automatic changes in green time durations. The system employs TensorFlow Lite and OpenCV on Raspberry Pi for processing video footage from the camera in real time, allowing for real-time changes in the timer. It successfully processed each frame in 0.2 s. The green time duration was varied from 10 to 30 s based on the number of vehicles, enhancing overall traffic. There is an approximate 50% reduction in waiting time and 37–43% reduction in queue length with an increase in traffic throughput by 10–15%. By optimizing traffic signal times based on actual conditions, the system was able to present an efficient, scalable and smart approach to traffic management.

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

Adaptive Traffic Light Control System Using Machine Learning

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Vaishali Savale,
  • Raj Bapat,
  • Sanskar Kulkarni,
  • Sanyam Kothari,
  • Rriddhi Rathi,
  • Roshan Raut

摘要

Traffic congestion in urban areas is an increasing challenge because of fixed traffic signals, which do not change according to real conditions. This paper brings in an Adaptive Traffic Light Control System based on computer intelligence and vision, allowing for automatic changes in green time durations. The system employs TensorFlow Lite and OpenCV on Raspberry Pi for processing video footage from the camera in real time, allowing for real-time changes in the timer. It successfully processed each frame in 0.2 s. The green time duration was varied from 10 to 30 s based on the number of vehicles, enhancing overall traffic. There is an approximate 50% reduction in waiting time and 37–43% reduction in queue length with an increase in traffic throughput by 10–15%. By optimizing traffic signal times based on actual conditions, the system was able to present an efficient, scalable and smart approach to traffic management.