Smart Traffic Control Using Adaptive Neuro-Fuzzy Inference System(ANFIS)

Abstract

Use of automobiles is increasing throughout the world, particularly in large urban areas. There has been the
problem of traffic congestion with the increase in the number of vehicles in cities. Thus, there is a requirement for smart
traffic control methods for better accommodating the increasing demand. Therefore, the transportation system will
continue to grow, and intelligent traffic controls have to be developed to face the road traffic problems. In this paper a
comparison has been drawn for Neuro-Fuzzy (NF) based smart traffic control system and fuzzy logic based smart traffic
control system. An adaptive neuro-fuzzy inference system is developed and tested against various traffic situations. Here
we have trained the Adaptive Neuro-Fuzzy Inference System (ANFIS) system by various traffic situations so that ANFIS
can draw the membership functions and corresponding rules by its own. Inputs which are generally used are Gap between
two vehicles, last time vehicles that haven’t passed during last green phase, delay at intersections, vehicle density, arrival
rate, leaving rate and queue length. Arrival rate of the particular phase and last time vehicles that haven’t passed during
last green phase are taken as inputs to the system by the considering the practical applicability. Output of both the system
are compared in terms of the average waiting time of the vehicles. ANFIS based traffic control system has been found
more efficient as the average delay of the vehicle and the waiting time of the vehicles at the int ersection have been
reduced