This study introduces an intelligent model for predicting urban traffic flow using an Adaptive Neuro-Fuzzy Inference System (ANFIS) integrated with chaotic behavior analysis through Lyapunov exponents. The primary objective is to address the challenges of nonlinear and dynamic traffic patterns by accurately modeling and forecasting sudden changes in traffic flow. The ANFIS model leverages the strengths of neural networks and fuzzy logic, providing a robust approach to capturing complex traffic behaviors. Chaotic behavior analysis, facilitated by Lyapunov exponents, enables the identification of critical and sensitive points in traffic systems, enhancing the predictive accuracy.



Real-world traffic data from Ferdowsi Boulevard in Mashhad, Iran, was used for training and testing the model. Data preprocessing included cleaning, normalization, and removal of irrelevant or inconsistent entries. The results demonstrated that the proposed model achieved an average prediction accuracy of 92% and significantly outperformed traditional statistical models such as ARIMA in detecting abrupt changes and chaotic behaviors.



This hybrid approach provides a powerful tool for urban traffic management, particularly in highly congested areas. The findings indicate that combining ANFIS with chaotic behavior analysis is effective for managing traffic flow and reducing congestion in urban environments. This methodology has the potential to be extended to other cities and regions, contributing to improved traffic forecasting and decision-making systems