Urban Traffic Fuzzy Prototypes Using a Graph- Based Two-Stage Clustering Algorithm

The problem of traffic congestion in both motorways and urban areas is getting worse every day. Therefore, new solutions must be found which are able to change the traffic situation efficiently. Fuzzy control theory introduces a rule-based approach to control systems with complicated behavior. The scheme, in fact, has proved to work well with problems such as the one we face here. Analyzing traffic data during different time intervals, clustering them into similar groups and finally extracting the traffic patterns as some fuzzy sets are necessary to study a traffic network system. The current research uses the agent-based model of an urban traffic network to gather data and construct the corresponding origin-destination (OD) matrices. A two-stage clustering algorithm (based on a fuzzy graph approach) is implemented in the next step. The traffic patterns, also known as mobility patterns, are extracted finally so as to provide the requirements for a fuzzy control system to be applied. The results are promising which confirms the validity of the modifications made in the method used for clustering.