Proposed Congestion Warning System Using YOLOv8 and DeepSORT model

The article proposes a system to identify traff c congestion in urban areas intending to be able to warn traff c jams directly to drivers and vehicles. Our method is to detect using YOLOv8 and DeepSORT to detect vehicles and assign IDs to each object. The density of vehicle traff c is calculated based on the zoned area and the average moving speed. We will determine the appropriate threshold to give the output to the system to notify the user about the situation based on the above-calculated parameters as well as the results in practice. Traff c status in the zoned area is informed by the telegram chat application. The system results have an accuracy of up to 92.9% and apply to larger models such as providing traff c information including road conditions, congested areas, travel time, and calculating traff c to regulate light signals to optimize traff c f ow.