Using Hybrid Neural Networks to Improve Traffic Prediction and Congestion Management

Urban regions are commonly plagued by traffic congestion, which results in substantial economic losses and a diminished quality of life. Accurate prediction of traffic flow and effective management of congestion are important in reducing the impacts of traffic. This paper presents a new approach using hybrid neural network models to enhance the accuracy of traffic predictions and improve strategies for congestion management. The proposed Materials and methods integrates Diffusion Convolutional Recurrent Neural Network (DCRNN) with graph-based models, allowing information to be shared among related sensors over large distances. The METR-Los Angeles (METR-LA) dataset consists of traffic data collected from 207 loop detectors located on highways in Los Angeles. Validation is done through various methods that prove the practicality and efficiency of the developed deep learning methodologies for real-time congestion monitoring and management systems.