Overview

Traffic forecasting estimates how movement and demand will change across road networks, transit systems, airspace, and communications networks, supporting planning, routing, and resource allocation. Common approaches combine graph neural networks, temporal convolutions, attention-based Transformers, and newer state-space models; other work uses signal decomposition, physical or agent-based simulation, and contextual data such as incidents, weather, or mobility patterns. Researchers are extending these models with probabilistic predictions, foundation models, and online, continual, or federated adaptation to handle uncertainty, privacy, and changing conditions. A central challenge is making forecasts accurate and efficient while ensuring they transfer across places, sensor configurations, and traffic regimes—and that improvements in forecast metrics actually help downstream decisions.

Latest papers

CardsList

92 papers7 papers in the last four weeks, down 30% on the four weeks before. 0.1% of all new papers.

  1. Efficient Traffic State Prediction With Dynamic Joint Spatio-Temporal Relation Inference

    Apr 10, 2025Zhifeng Hao, Kai Hu, Juncai Zhang +2Temporal Convolutional NetworksDynamic Graph Learning

  2. Spatio-Temporal Partial Sensing Forecast for Long-term Traffic

    Aug 2, 2024Zibo Liu, Zhe Jiang, Zelin Xu +5Long-Term Time Series ForecastingSpatiotemporal Forecasting