Real-Time Traffic Light Optimization Using Yolov9 and Length-Based Metrics

The Indian traffic control system faces lots of difficulties due to the increasing volume of vehicles, ineffective systems for traffic administration during peak hours, and the frequent need for manual intervention due to the inadequate performance of traffic signals in managing heavy traffic flow. Traditional traffic lights in India have defined timings for each lane, which frequently cause longer traffic jams in lanes with more traffic. This study presents an intelligent traffic control system that incorporates the YOLOv9 model for real-time traffic length prediction and intelligently allocates green, red, and orange signal timings. YOLOv9 builds a bounding box that allows it to compute vehicle density precisely by enclosing the initial and final cars in every frame. As a comparison with traditional fixed-time methods, the proposed approach recalculates traffic signal timings at each cycle based on the recorded length of traffic. In order to ensure that emergency services respond more quickly, the system efficiently prioritizes emergency vehicles by quickly moving them from their lane when they are spotted. This adaptive strategy aligns signal duration with real traffic demand, boosting overall traffic efficiency and regulating traffic flow by reducing unnecessary wait times for low-traffic lanes. In comparison to fixed-timing systems and object detection strategies, our research on adaptive traffic systems reveals a 66% to 77% decrease in vehicle delay. Compared to traditional fixed-timing approaches, the proposed method demonstrates significant enhancements in effectively managing traffic congestion.