The swift growth of the population in the various urban centers has led to the over-dependence of people on the use of cars, thereby causing traffic congestion in cities, which seems to be alarming. It is essential for cities to satisfy the need for efficient mobility solutions as a minimum of half of the population belongs to cities according to a survey. The traditional systems used to manage the traffic usually involve static infrastructures and manual control which make it hard for them to solve the traffic problems. The proposed work sets forth a methodology wherein a Modified YOLO (You Only Look Once) object detection is coupled with AI, computer vision, and IoT to make traffic management effective. The surveillance system works in real time and identifies and tracks vehicles, people, or any other urban objects through smart sensors. Major improvements include enhanced detection accuracy, scalability in the ability to handle increasing urban data volume, and lessened computational complexity incorporated by edge-based architectures. It is depicted from the experimental results that the framework cum methodology proposed can perform effective object detection with its accuracy at 90%, decreases the traffic flow by 23 and enhances vehicle travel time by 15.