Automated Traffic Signal Controlling Using Deep Learning Techniques
The present Traffic Control Systems (TCS) in the metro urban communities of India is
inefficient due to arbitrariness in the rush hour at crossroads throughout the day. A run of the mill day
in India would take a gander at peak hour timings when the traffic thickness is high in the streets and
peak hour timings when the traffic thickness isn't so high. The traffic flags at all the intersections in
India are hardcoded which means the signals have fixed memories and switch traffic between statically.
Because of this, the vehicles need to wait for prolonged amount of time even though the traffic density
is less. The solution to this issue is by building up a framework which distinguishes traffic densities on
each lane of the junction switch the signal lights dynamically along with synchronization of the adjacent
lane’s traffic signal. This process will be divided into two modules. The first module will comprise of
building a model which will detect and the number of vehicles in particular lane were counted. All these
lane’s density will be added in a dynamic queue on which pre-emptive and dynamic scheduling
algorithm will be applied.