Traffic Datasets

Traffic datasets are crucial for developing and evaluating intelligent transportation systems, aiming to improve traffic prediction, management, and safety. Current research focuses on addressing the spatio-temporal heterogeneity and incompleteness of traffic data using advanced deep learning models, such as graph neural networks, transformers, and recurrent neural networks, often incorporating attention mechanisms and other techniques to enhance accuracy and efficiency. These advancements are significant for improving urban planning, optimizing resource allocation, and enabling more effective real-time traffic control and autonomous driving systems. The development and sharing of large, high-quality, and diverse traffic datasets, including those incorporating event-based data and 3D LiDAR information, are also key areas of ongoing work.