DMSTG: Dynamic Multiview Spatio-Temporal Networks for Traffic Forecasting
DMSTG: Dynamic Multiview Spatio-Temporal Networks for Traffic Forecasting
复制标题
DOI：
10.1109/tmc.2023.3328038
复制
发表时间：
2024-06
影响因子：
7.9
通讯作者：
Zulong Diao;Xin Wang;Dafang Zhang;Gaogang Xie;Jianguo Chen;Changhua Pei;Xuying Meng;Kun Xie-Kun-Xi
中科院分区：
文献类型：
--
作者：
Zulong Diao;Xin Wang;Dafang Zhang;Gaogang Xie;Jianguo Chen;Changhua Pei;Xuying Meng;Kun Xie-Kun-Xi
交通传感器网络被广泛应用于智慧城市中，用于实时监控交通。利用这些数据来预测未来的交通状况，有可能提高智能交通系统的决策能力，这引起了工业界和学术界的广泛关注。其中，基于图卷积神经网络（GCN）的全网络预测已成为主流。它用一个预定义的拉普拉斯矩阵在图中对传感器的空间依赖性进行建模。然而，由于不同时期或不同区域的交通模式存在巨大差异，因此了解时空交通模式是一项相当具有挑战性的工作。此外，由于恶劣的通信条件或传感器故障导致不可避免的数据丢失，实际收集的数据可能会受到污染。针对这些问题，本文提出了一种考虑局部/全局、短期/长期时空依赖及其动态变化等因素的动态多视角时空预测框架。我们创造性地设计了两个不同的模块，以全面感知交通模式的变化。我们首先提出了一个基于理论推导的动态学习模块，用于实时估计GCN图的拉普拉斯矩阵。我们还设计了一个基于自关注的模块来动态地为交通数据中的各个部分分配权重。通过特征融合模块对多个视角的时空特征进行深度融合。用5个实时交通数据集对预测效果进行了评价。实验结果表明，我们的框架可以始终优于最先进的基线，并且在噪声环境下具有更强的鲁棒性。
Traffic sensor networks are widely applied in smart cities to monitor traffic in real-time. Exploiting such data to forecast future traffic conditions has the potential to enhance the decision-making capabilities of intelligent transportation systems, which attracts widespread attention from both industries and academia. Among them, network-wide prediction based on graph convolutional neural networks(GCN) has become mainstream. It models the spatial dependencies of sensors in a graph with a pre-defined Laplacian matrix. However, understanding spatio-temporal traffic patterns is quite challenging as there is a huge difference in terms of traffic patterns during different periods or in different regions. In addition, the actual data collected can be polluted due to unavoidable data loss from severe communication conditions or sensor failures. Considering these issues, we propose a novel dynamic multiview spatial-temporal prediction framework which takes into consideration various factors, including local/global, short/long term spatio-temporal dependencies and their dynamic changes. We creatively design two different modules to comprehensively perceive the changes in traffic patterns. We first propose a dynamic learning module based on our theoretical derivation to estimate the Laplacian matrix of the graph for GCN timely. We also design a self-attention based module to dynamically assign a weight to each part in traffic data. The spatio-temporal features from multiple views are deeply fused by a feature fusion module. The forecasting performance is evaluated with 5 real-time traffic datasets. Experiment results demonstrate that our framework can consistently outperform the state-of-the-art baselines and be more robust under noisy environments.