AI Traffic Control System Based on Deepstream and IoT Using NVIDIA Jetson Nano

With the rapid growth of population and economic development, the number of vehicles is increasing day by day. Which leads to traffic congestion. In certain time slots of the day, it becomes unbearable. So, it has become very difficult to control the traffic and maintain traffic law and order traditionally. Thus, it creates a demand for an improved and optimized traffic management system. Fixed timers or manual control are being used in conventional ways of traffic control, which has been proven ineffective to mitigate this situation. In this paper, a completely new system has been introduced to solve the problem of traffic congestion. Which will monitor vehicle traffic in real-time and the system is smart enough to detect traffic conditions and provide decisions depending on the situation. Besides that, the system is designed in a such way, that can also detect the number plate of those vehicles which violate the traffic rules. Advanced features have been added to inform law enforcement agencies about that number plate and send a confirmation message to the driver mentioning the amount of penalty. Throughout the paper, attempts have been made to represent this system's advancement and efficiency over the existing manual traffic control system.