Pattern recognition method for simplified coordinated traffic signal control

Abstract

As traffic conditions on arterial streets vary by day and month, fixed time signal control do not cope optimally with traffic demand. Especially megacities suffering from traffic congestions would need traffic responsive control systems. In this paper a low-cost real-time coordinated traffic signal control system is presented which is suited in particular for megacities in developing or newly industrializing countries. The proposed system is based on a pattern recognition method using feed forward artificial neural networks. The system recognizes the traffic state by system detectors and selects suitable settings out of a set of previously optimized timing plans. Real-time software in the loop simulation technique is used to compare the proposed system with optimized fixed time signal control. Although this study is based on experiments with microscopic traffic flow simulation, the algorithm itself is ready for real world implementations.

Integrated expression method for technical mapping of traffic parameters using RGB color model

“…In order to predict future traffic flows, numerical approaches have been widely used . These researchers have tried to use the Greenshield model for the purpose of prediction.…”

“…Genetic and neural network algorithms were applied in pattern analysis of traffic flow. Chung conducted pattern analyses using simple matching methods for the similarity of traffic scenarios.…”