caldenrodrigues / traffic-signal-duration-control Goto Github PK
View Code? Open in Web Editor NEWDevelopment of Traffic Signal Duration Control using Deep Learning is the title of our project. Our system will monitor a traffic junction. It will collect the video data and send it to the cloud. It will then detect the number of vehicles, the direction of the vehicle and the type of vehicle. By applying an algorithm on the data received it will suggest the most ideal timing to switch between Green to Red or Vice-Versa in real time. This system will thereby reduce the traffic flow in real time. Our system will also make prediction based on previous data of congestion. It will take data from the previous week to make sure frequently happening congestions are avoided. Our system will also find patterns and will predict if a congestion can occur by using previously available data of traffic hence suggesting the most ideal timing to switch between Green to Red or Vice Versa in real time. It will also predict the traffic signal duration based on the information gained from the peer signals hence creating a massive grid which is controlled by central unit. Each signal can inform the adjacent signal about the outgoing traffic from its area. This can help the adjacent signals to take appropriate measures to avoid congestion hence reducing traffic flow and creating an optimal go green time for a user.