Abstract:
Traffic congestion is a very common and a big issue nowadays in different countries
around the world. The intelligent/smart traffic signaling system is crucial for an
effective flow of traffic and has got much attention from researchers. Over the lastfew
years, some intelligent systems have been developed using different techniques. In
Pakistan, the current traffic signaling is still handcrafted. Traffic congestion is also a
cumbersome task for traffic wardens. They set the time for each signal manually. In
this regard, this project aims at developing an adaptive traffic signaling system for
urban areas that will dynamically adjust signal timings according to the density of
traffic. Our intelligent system will take videos of each side of intersection and extract
frames. It will generate vehicle on simulation that will provide the data further to a
deep reinforcement learning model which assigns the time to a signal, depending upon
the ratio of vehicles and their waiting time on each side of the intersections dynamically.
In this way, traffic congestion at the intersection will be reduced in an effective and
efficient manner.
