Urban traffic fuzzy prototypes using a graph-based two-stage clustering algorithm

Abstract

The problem of traffic congestion in both motorways and urban areas is getting worse every day. Therefore some new solutions must be found which are potentially able to change the traffic situation efficiently. Fuzzy control approach is a rule-based methodology for controlling systems with complicated behavior. The approach, however, has proved to work well with problems such as the one we face here. Studying traffic data during different time intervals and clustering them into similar groups and finally extracting the traffic patterns in the form of some fuzzy sets are necessary for this study. The current study uses an agent-based modeling of an urban traffic network in order to gather data and construct the corresponding matrices. Then a two-stage clustering algorithm based on a fuzzy graph approach is implemented and the mobility patterns are extracted finally in order to provide the needs for a fuzzy control system to be applied. The results show a promising sight of the alterations made for the methodology used for clustering.