Abstract

Urban traffic congestion has become a critical issue that not only affects the quality of daily lives but also harms the environment and economy. traffic patterns are recurrent in nature, so is congestion. However, little attention has been paid to the development of methods that would enable early warning of the formation of congestion and its propagation. This paper proposes a method for automated early congestion detection operating over time horizons ranging from half an hour to three hours. The method uses a deep learning technique, Convolutional Neural Networks (CNN), and adapts it to the specific context of urban roads. Empirical results are reported from a busy traffic corridor in the city of Bath. Comprehensive evaluation metrics, including Detection Rate, False Positive Rate and Mean Time to Detection, are used to evaluate the performance of the proposed method compared to more conventional machine learning methods including Feed-forward Neural Network and Random Forest. The results indicate that recurrent congestion can indeed be predicted before it occurs and demonstrates that CNN based method offers superior detection accuracy compared to the conventional machine learning methods in this context.

Abstract

Predicting traffic conditions from online route queries is a challenging task as there are many complicated interactions over the roads and crowds involved. In this paper, we intend to improve traffic prediction by appropriate integration of three kinds of implicit but essential factors encoded in auxiliary information. We do this within an encoder-decoder sequence learning framework that integrates the following data: 1) offline geographical and social attributes. For example, the geographical structure of roads or public social events such as national celebrations; 2) road intersection information. In general, traffic congestion occurs at major junctions; 3) online crowd queries. For example, when many online queries issued for the same destination due to a public performance, the traffic around the destination will potentially become heavier at this location after a while. Qualitative and quantitative experiments on a real-world dataset from Baidu have demonstrated the effectiveness of our framework. Comment: KDD 2018. The first two authors share equal contribution