Abstract:Although traffic prediction has been receiving considerable attention with a number of successes in the context of intelligent transportation systems, the prediction of traffic states over a complex transportation network that contains different road types has remained a challenge. This study proposes a multi-scale graph wavelet temporal convolution network (MSGWTCN) to predict the traffic states in complex transportation networks. Specifically, a multi-scale spatial block is designed to simultaneously capture the spatial information at different levels, and the gated temporal convolution network is employed to extract the temporal dependencies of the data. The model jointly learns to mount multiple levels of the spatial interactions by stacking graph wavelets with different scales. Two real-world datasets are used in this study to investigate the model performance, including a highway network in Seattle and a dense road network of Manhattan in New York City. Experiment results show that the proposed model outperforms other baseline models. Furthermore, different scales of graph wavelets are found to be effective in extracting local, intermediate and global information at the same time and thus enable the model to learn a complex transportation network topology with various types of road segments. By carefully customizing the scales of wavelets, the model is able to improve the prediction performance and better adapt to different network configurations.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is traffic state prediction in complex traffic networks. Specifically, existing traffic prediction methods face challenges when dealing with complex traffic networks containing different types of roads. These problems include: 1. **Multi - scale extraction of spatial information**: Traditional graph convolutional networks (GCNs) assume that the interactions between nodes and their neighboring nodes are uniform, which is not always true in the real world. Moreover, spatial features are often concentrated on the main nodes and their neighbor nodes within a limited range, which may not reflect the actual situation because the scale of traffic propagation may change due to external factors (such as accidents, bad weather, construction, etc.). Therefore, a method that can capture local, intermediate, and global information simultaneously is needed to better represent the traffic state. 2. **Modeling of time - dependence**: Traffic data usually has a strong time - correlation, that is, the current traffic state may be affected by previous traffic conditions. Although recurrent neural networks (RNNs) and their variants (such as LSTM and GRU) are used to capture time - dependence, they may encounter the problem of gradient explosion when dealing with long - sequence data. In contrast, gated time - convolutional networks (TCNs) process sequence data in a non - recursive way, which can alleviate this problem and support parallel computing to improve efficiency. 3. **Different types of road segments**: Complex traffic networks are composed of multiple types of road segments (such as highways, ramps, arterial roads, one - way streets, etc.). How to effectively extract the spatial features of these different types of road segments in one model is a challenge. Existing GCN - based methods may be insufficiently sensitive to different network structures and difficult to achieve effective traffic prediction. To address the above challenges, this paper proposes a multi - scale graph wavelet time - convolutional network (MSGWTCN), aiming to simultaneously learn spatial and temporal features at multiple levels, thereby improving the accuracy of traffic state prediction. This model combines the multi - scale graph wavelet network (MS - GWN) and the gated time - convolutional network (TCN) and can effectively extract and utilize multi - scale structural information in complex traffic networks.

Traffic Prediction considering Multiple Levels of Spatial-temporal Information: A Multi-scale Graph Wavelet-based Approach

A Spatial-Temporal Gated Hypergraph Convolution Network for Traffic Prediction

Spatial–Temporal Deep Tensor Neural Networks for Large-Scale Urban Network Speed Prediction

Lane-Level Traffic Flow Prediction with Heterogeneous Data and Dynamic Graphs

Towards City-Scale Traffic Prediction: Cluster Graph Wavelet Networks.

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

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

Wavelet-Inspired Multiscale Graph Convolutional Recurrent Network for Traffic Forecasting

Channel attention-based spatial-temporal graph neural networks for traffic prediction

Multistep traffic speed prediction from multiple time-scale spatiotemporal features using graph attention network

A Spatiotemporal Graph Wavelet Neural Network for Traffic Flow Prediction

Multi-View Spatial–Temporal Graph Convolutional Network for Traffic Prediction

Temporal Multi-Graph Convolutional Network for Traffic Flow Prediction

MST-GNN: graph neural network with multi-granularity in space and time for traffic prediction

Causally-Aware Spatio-Temporal Multi-Graph Convolution Network for Accurate and Reliable Traffic Prediction

Traffic Flow Prediction via Spatial Temporal Graph Neural Network

A Multi-Scale Residual Graph Convolution Network with hierarchical attention for predicting traffic flow in urban mobility

Traffic flow prediction based on temporal multigraph convolutional neural network

Multi-Head Spatiotemporal Attention Graph Convolutional Network for Traffic Prediction

Citywide Traffic Flow Prediction Based on Multiple Gated Spatio-temporal Convolutional Neural Networks

Dual-channel Spatio-Temporal Wavelet Transform Graph Neural Network for Traffic Forecasting.