Spatiotemporal Recurrent Convolutional Networks for Traffic Prediction in Transportation Networks

Yu, Haiyang; Wu, Zhen; Wang, Shuqin; Wang, Yunpeng; Ma, Xiaolei

doi:10.3390/s17071501

Cited by 488 publications

(250 citation statements)

References 44 publications

(57 reference statements)

Supporting

Mentioning

250

Contrasting

Order By: Relevance

“…We compared the proposed model with five baseline deep NN models, namely, LSTMs (Hochreiter and Schmidhuber, 1997), NLSTM (Moniz and Krueger, 2018), DCNNs , CapsNet (Sabour et al, 2017), and CNN+LSTMs (Yu et al, 2017), to evaluate its prediction performance. The details of the CNN+LSTM baseline model is shown in…”

Section: Baseline Modelsmentioning

confidence: 99%

“…However, the actual structure of a complicated roadway network cannot be properly represented by a 2D matrix, and CNNs inevitably capture a certain amount of spurious spatial relationships. The second strategy (Zhang et al, 2016;Yu et al, 2017) employs CNNs to capture the spatial dependencies by projecting various traffic states to their corresponding physical roadway links using different colors and by processing a traffic network map as an image. In this way, the actual spatial features of the traffic network are learned.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Forecasting Transportation Network Speed Using Deep Capsule Networks With Nested LSTM Models

Zhong

et al. 2021

IEEE Trans. Intell. Transport. Syst.

Self Cite

View full text Add to dashboard Cite

Accurate and reliable traffic forecasting for complicated transportation networks is of vital importance to modern transportation management. The complicated spatial dependencies of roadway links and the dynamic temporal patterns of traffic states make it particularly challenging. To address these challenges, we propose a new capsule network (CapsNet) to extract the spatial features of traffic networks and utilize a nested LSTM (NLSTM) structure to capture the hierarchical temporal dependencies in traffic sequence data. A framework for network-level traffic forecasting is also proposed by sequentially connecting CapsNet and NLSTM. On the basis of literature review, our study is the first to adopt CapsNet and NLSTM in the field of traffic forecasting. An experiment on a Beijing transportation network with 278 links shows that the proposed framework with the capability of capturing complicated spatiotemporal traffic patterns outperforms multiple state-of-the-art traffic forecasting baseline models. The superiority and feasibility of CapsNet and NLSTM are also demonstrated, respectively, by visualizing and quantitatively evaluating the experimental results.

show abstract

Section: Baseline Modelsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Forecasting Transportation Network Speed Using Deep Capsule Networks With Nested LSTM Models

Zhong

et al. 2021

IEEE Trans. Intell. Transport. Syst.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Zhang et al [26] designed an architecture including attention mechanism, GCN and sequence-to-sequence model to conduct multistep speed prediction. After ConvLSTM was firstly introduced [27], CNN and LSTM are often integrated together to perform traffic predictions [28,29]. Recently, generative adversarial network has began to attract researchers' attention and has been applied to traffic time estimation [30].…”

Section: Introductionmentioning

confidence: 99%

Cluster-Based LSTM Network for Short-Term Passenger Flow Forecasting in Urban Rail Transit

2019

View full text Add to dashboard Cite

Short-term passenger flow forecasting is a crucial task in the operation of urban rail transit. Emerging deeplearning technologies have become effective methods to overcome this problem. In this study, we propose a deep-learning architecture called Conv-GCN combining graph convolutional network (GCN) and 3D convolutional neural network (3D CNN). First, we introduce a multi-graph GCN to deal with three patterns (recent, daily, and weekly patterns) of inflow and outflow separately. Multi-graph GCN network can capture spatiotemporal correlations and topological information in a whole network. Then, a 3D CNN is applied to deeply integrate the inflow and outflow information. High-level spatiotemporal features between different patterns of inflow and outflow, and between stations nearby and far away can be extracted by 3D CNN. Finally, a fully connected layer is used to output results. The Conv-GCN model is evaluated on smart card data of Beijing subway under the time interval of 10 min, 15 min, and 30 min. Results show that this model performs the best among seven other relative models. In terms of the RMSE, the performances under three time intervals have been improved by 9.402%, 7.756%, and 9.256%, respectively. This study can provide critical insights for subway operators to optimize the operation.

show abstract

“…Several publications have developed complex neural network or deep learning models for traffic state prediction and, to a lesser extent, missing data imputation [25], [26], [27], [28]. Such methods have been shown to provide accurate predictions, especially for complex, high dimensional traffic data, though no previous work has applied this class of models to the task of probe vehicle-based link-level speed or travel time data completion.…”

Section: Introductionmentioning

confidence: 99%

Multi-Output Gaussian Processes for Crowdsourced Traffic Data Imputation

Rodrigues

Henrickson

Pereira

2019

IEEE Trans. Intell. Transport. Syst.

View full text Add to dashboard Cite

Traffic speed data imputation is a fundamental challenge for data-driven transport analysis. In recent years, with the ubiquity of GPS-enabled devices and the widespread use of crowdsourcing alternatives for the collection of traffic data, transportation professionals increasingly look to such usergenerated data for many analysis, planning, and decision support applications. However, due to the mechanics of the data collection process, crowdsourced traffic data such as probe-vehicle data is highly prone to missing observations, making accurate imputation crucial for the success of any application that makes use of that type of data.In this article, we propose the use of multi-output Gaussian processes (GPs) to model the complex spatial and temporal patterns in crowdsourced traffic data. While the Bayesian nonparametric formalism of GPs allows us to model observation uncertainty, the multi-output extension based on convolution processes effectively enables us to capture complex spatial dependencies between nearby road segments. Using 6 months of crowdsourced traffic speed data or "probe vehicle data" for several locations in Copenhagen, the proposed approach is empirically shown to significantly outperform popular state-ofthe-art imputation methods.

show abstract

Spatiotemporal Recurrent Convolutional Networks for Traffic Prediction in Transportation Networks

Cited by 488 publications

References 44 publications

Forecasting Transportation Network Speed Using Deep Capsule Networks With Nested LSTM Models

Forecasting Transportation Network Speed Using Deep Capsule Networks With Nested LSTM Models

Cluster-Based LSTM Network for Short-Term Passenger Flow Forecasting in Urban Rail Transit

Multi-Output Gaussian Processes for Crowdsourced Traffic Data Imputation

Contact Info

Product

Resources

About