Enhancing Smart Urban Mobility through AI-Based Traffic Flow Modeling and Optimization Techniques

Abstract

The growing challenges in city traffic warrant smart methods to relieve congestion and improve both traffic organization and sustainability. This work explores using Artificial Intelligence methods such as machine, deep and reinforcement learning, to help in modeling and improving traffic flow in smart cities. The traffic prediction task is achieving with LSTM networks and optimizing traffic signals is done with DQNs using real-time traffic information from cities. According to the experiments, traffic time, congestion and CO₂ levels are all lower by 18%, 25% and 12% respectively, when compared with traditional systems. Applying AI to transportation helps ITS, ensuring the city’s transportation systems are accessible, targeted to users’ needs, eco-friendly and improved.