Smart Signals: Real-Time Traffic Adjuster

The omnipresent problem of traffic congestion in cities, even those well equipped with traffic lights and cameras, causes daily trouble to one person or another. In this research paper, we have created a better traffic management system using a object detection-based and time series analyzer-based traffic light controller that can automatically adjust to the volume of vehicles at the traffic signal in order to reduce congestion. Hence, a system to reduce wait times and prevent traffic jams is required. The proposed methodology predicts the expected traffic intensity, and detects the real-time traffic intensity; and thus makes sure that the direction with more traffic receives a green light for a longer period of time in an appropriate proportion compared to the way with less traffic. This approach utilises YOLOv2 for image detction of vehicles - including cars, trucks and bikes. This has resulted in almost 33% savings in terms of wait times, and thus will enable traffic to move efficiently through intersections without human interference and will also aid in minimizing pollution.