An Intelligent GIS Based Physical Model for Improvement of Urban Traffic System

Abstract

This is a thesis on a system which will be able to sense the traffic situation of a road; and with its prior knowledge of neural network it will be able to predict the possibility of traffic congestion so that a better decision can be taken by the authority. The number of vehicles coming from residential, industrial or restricted area will also be taken into consideration to get a better response at the time of prediction. With the help of neural network a model for the entire possible scenario of congestion is trained and this learning has helped to provide the output scenario for any given situation. The output of the system will be the percentage occupation of the road or congestion depending on which authority will take the wise decision to remove the congestion. This can also compare the current situation of different roads to learn the percentage of congestion. Then if the information of the traffic situation is made available through internet then general people can get the idea of congestion and avoid the busy roads to select an alternative free one. Government agencies will be able to take decisions like where to construct new roads, install traffic lights etc. if there is authentic information regarding the behavior of traffic in different parts of the day. Gathering of this information is easily possible from our project