Tumgik
fukatsoft-blog · 5 years
Text
Why Traffic Signal Automation is Important?
Background:
With the increase in numbers of traffic vehicles and not well- structured flow of traffic in the populated countries like Pakistan. There is a significant loss of time and ultimately money. So, to control the bulk of traffic and unstructured flow of data traffic signals were introduced. It really controlled the traffic and by introducing the e-challan all the citizens used to follow the traffic rules but the time for the signals change is fixed that is the traffic has to wait for the fixed amount of time so it causes some major problems that are as follows
  There may be an ambulance waiting for the green signal for fixed amount of time.
  There might be a huge number of traffic vehicles on one signal and there is no traffic on other side of the signal.
  Excessive delays due to fixed traffic signal time without any traffic on other side of the signal can translate to wasted fuel and air pollution. 
  Increase in Rear-End Collisions due to large volume of traffic on one side of the signal.
Why we need signal automation:  
We have explained some disadvantages of traffic signals with fixed amount of time delays. So, there is need of a Signal automated system that detects the number of traffic vehicles and gives more time to the bulk of the traffic to flow though the signal. Similarly, the system will allow the traffic to low first where there is an ambulance or fire brigade. There is also the need of the system that will allow the government authorities to go first with more amount of time given to that side of signal.
With the help of Artificial Intelligence techniques and very high performance GPUs it would be feasible to make such Automated Systems for Traffic and road Safety. As the field of deep learning is improving and getting famous  quickly certain automated systems have been developed for last some years to improve the flow of traffic. So ,we are doing our work to develop a system to overcome the disadvantages of traffic signal defined above. It is also explained in our blog how does it help.
Our strategy:
We will use the following techniques to develop and deploy this automated system.
  We will detect ambulance or fire brigade from the traffic and will give more time to that side of the signal which contain these vehicles.
  We will count the total number of vehicles on each side of the signal and will give the side more time with most number of vehicles
  We will give more time to that side of the signal where any government high authority is present.
Our tools and technologies
We will use convolution neural network model for this task. we will use Kerala library for building our neural network with tensor-flow as back-end and use the python programming language. We also need high speed GPUs for training of our model.
1 note · View note