Abstract:Traffic prediction is a core problem in the intelligent transportation system and has broad applications in the transportation management and planning, and the main challenge of this field is how to efficiently explore the spatial and temporal information of traffic data. Recently, various deep learning methods, such as convolution neural network (CNN), have shown promising performance in traffic prediction. However, it samples traffic data in regular grids as the input of CNN, thus it destroys the spatial structure of the road network. In this paper, we introduce a graph network and propose an optimized graph convolution recurrent neural network for traffic prediction, in which the spatial information of the road network is represented as a graph. Additionally, distinguishing with most current methods using a simple and empirical spatial graph, the proposed method learns an optimized graph through a data-driven way in the training phase, which reveals the latent relationship among the road segments from the traffic data. Lastly, the proposed method is evaluated on three real-world case studies, and the experimental results show that the proposed method outperforms state-of-the-art traffic prediction methods.

What problem does this paper attempt to address?

This paper mainly explores the problem of traffic flow prediction, which is an important issue in intelligent transportation systems and has wide applications in traffic management and planning. The current challenge is how to effectively mine the spatial and temporal information of traffic data. The paper proposes an optimized graph convolution recurrent neural network (OGCRNN) to solve this problem. Traditional deep learning methods, such as convolutional neural networks (CNN), have shown good performance in traffic prediction. However, they sample traffic data into regular grids, which disrupts the natural spatial structure of road networks. The paper introduces the concept of graph networks, constructs a graph to represent the spatial information of road networks, and the proposed method learns to optimize the graph in a data-driven manner during the training phase, revealing the potential relationships between road segments in the time series data of road segments. Experimental results show that the proposed OGCRNN method outperforms existing traffic prediction methods in three real case studies, demonstrating its superiority in predicting future 15-minute and 30-minute traffic flows. The main contributions of this method include constructing a graph-based spatio-temporal neural network, proposing a graph updating strategy, and constructing an optimized graph matrix. The paper also reviews related work, including statistical, machine learning, and deep learning-based traffic prediction methods, and provides a detailed introduction to the components of the proposed method, such as data representation, graph convolution, GRU network, and the working principle of OGCRNN. Finally, the paper provides experimental settings and comparison methods to validate the effectiveness of OGCRNN.

Optimized Graph Convolution Recurrent Neural Network for Traffic Prediction

A Spatial-Temporal Gated Hypergraph Convolution Network for Traffic Prediction

Traffic Graph Convolutional Recurrent Neural Network: A Deep Learning Framework for Network-Scale Traffic Learning and Forecasting

Dynamic multi-scale spatial-temporal graph convolutional network for traffic flow prediction

Dual Dynamic Spatial-Temporal Graph Convolution Network for Traffic Prediction

Dynamic Graph Convolutional Recurrent Network for Traffic Prediction: Benchmark and Solution

Graph Pruning Based Spatial and Temporal Graph Convolutional Network with Transfer Learning for Traffic Prediction

Multi-graph Spatio-temporal Graph Convolutional Network for Traffic Flow Prediction

Hierarchical Graph Convolution Network for Traffic Forecasting

Dynamic graph convolutional network for long-term traffic flow prediction with reinforcement learning

Predicting Traffic Flow Using Dynamic Spatial-Temporal Graph Convolution Networks

Traffic Flow Prediction via Spatial Temporal Graph Neural Network

Dynamic Graph Convolutional Network with Attention Fusion for Traffic Flow Prediction

A Graph Convolutional Method for Traffic Flow Prediction in Highway Network

T-GCN: A Temporal Graph ConvolutionalNetwork for Traffic Prediction

Graph learning-based spatial-temporal graph convolutional neural networks for traffic forecasting

Traffic flow prediction based on graph convolutional networks with a parallel attention network and stacked gate recurrent units

A deep learning traffic flow prediction framework based on multi-channel graph convolution

T-GCN: A Temporal Graph Convolutional Network for Traffic Prediction

Spatial-Temporal Tensor Graph Convolutional Network for Traffic Prediction

Spatial-Temporal Traffic Flow Prediction with Fusion Graph Convolution Network and Enhanced Gated Recurrent Units