Multi Layer Perception Type Artificial Neural Network based Traffic Control

Real time traffic control is a main criteria of the urban traffic signal control system, and giving viable ongoing traffic signal control for a substantial complex traffic system is a testing issue. The objective of the work is to find and adjust the timing of signals based on the traffic density. Such a situation arises in a city where outbound vehicles during morning time and inbound vehicles during evening time is more while the vehicular movement in the opposite direction is less. To predict and adjust the timings of the signals on both sides of the road at the same time, Artificial Neural Network technique is used. A real time traffic survey of Light Motor Vehicle, Heavy Motor Vehicle, two and three wheeler vehicular movement in Thanjavur city is done. The number of vehicles (cars, auto, bikes, trucks and buses) and width of the road was given as a input and the output predicted was in terms of timing for the traffic signal at any particular place and for any particular width of the road. The width of the road is also taken into account which is essential in planning a city based traffic consisting of different road widths.