INTELLIGENT TRAFFIC LIGHT CONTROL SYSTEM

Abstract

Traffic is a major problem in highly populated countries like India. The present traffic system assigns
signals for a fixed interval of time. This creates a problem when there are more vehicles on one side. Assigning a
green signal to an empty road is also of no use. The other problem is with emergency vehicles. Delay of these
vehicles may result in the loss of some lives. We are proposing a model to solve all these problems that
dynamically assigns signal timing based on vehicle density. This model gives priority to emergency vehicles. We
place a camera at the traffic signal to take images of all directions, which are sent to Raspberry Pi. These images
are processed using the most advanced vehicle density on each lane are provided via real-time object
identification technology dubbed YOLO ONLY LOOK ONCE (YOLO), which is based on a sophisticated
neural network. This reduces waiting time which intern reduces fuel wastage. In the future, we can extend this
model to multiple traffic lights and make a network. The information can be passed from one another and
congestion can be easily avoided by getting prior information.