Urbanization has intensified transportation challenges such as congestion, inefficiency, and safety risks. Traditional traffic management approaches often fail to adapt to dynamic urban conditions, leading to delays and environmental burdens. This paper explores how Machine Learning (ML)-driven Intelligent Transportation Systems (ITS) can optimize urban mobility by integrating predictive, adaptive, and data-driven solutions. Using models such as Long Short-Term Memory (LSTM) networks, Convolutional Neural Networks (CNNs), and reinforcement learning, ITS applications are analyzed in traffic flow prediction, adaptive signal control, smart parking, and AI-powered public transport scheduling. Real-world implementations in cities including Los Angeles, Barcelona, Singapore, and Dubai demonstrate the potential of ML-enabled ITS to reduce congestion, improve safety, and enhance sustainability. The paper also addresses critical challenges such as data privacy, infrastructure costs, and algorithmic fairness, while highlighting future research directions in federated learning, vehicle-to-everything (V2X) communication, and integration with autonomous vehicles. Findings underscore that ML-driven ITS is not only key to improving transportation efficiency but also a foundation for building sustainable smart cities.
Patil et al. (Thu,) studied this question.