Dynamic Origin-Destination Prediction in Urban Rail Systems: A Multi-Resolution Spatio-Temporal Deep Learning Approach

Noursalehi, Peyman; Koutsopoulos, Haris N.; Zhao, Jinhua

doi:10.1109/tits.2020.3047047

Cited by 62 publications

(21 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Specifically, Zhang et al [11] developed a channel-wise attentive split-convolutional neural network to assign different values for OD features, while Cheng et al [14] developed a high-order weighted dynamic mode decomposition to learn time-evolving features of a metro system. Recently, Noursalehi et al [13] toughly used the historical DO matrices to forecast the future OD matrices with a multi-resolution spatial-temporal neural network model. However, all previous works have never explicitly exploited the information of unfinished transactions.…”

Section: Traffic Origin-destination Predictionmentioning

confidence: 99%

“…Specifically, U t (i, j) denotes the number of passengers that entered at station i at time interval t but have not reached their destinations. Such information has never been explored in previous works [10], [12], [13].…”

Section: ) Transaction Datamentioning

confidence: 99%

“…In literature, there are very few methods [10], [11], [12], [13], [14] proposed for online metro origin-destination ridership prediction. Conventional works either take the incomplete historical OD matrices as input or directly utilize DO matrices to forecast the future OD matrices.…”

Section: Introductionmentioning

confidence: 99%

“…Therefore, more information is available for metro OD prediction. Second, it is sub-optimal to directly use the historical DO information to toughly forecast the future OD matrices [13], since the former isn't the essential factor affecting the evolution of the latter. Third, all abovementioned methods are unaware of the mutual information between OD and DO ridership, i.e., forecasting the future OD and DO matrices separately.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Online Metro Origin-Destination Prediction via Heterogeneous Information Aggregation

Liu¹,

Zhu²,

Li³

et al. 2021

Preprint

View full text Add to dashboard Cite

Metro origin-destination prediction is a crucial yet challenging time-series analysis task in intelligent transportation systems, which aims to accurately forecast two specific types of cross-station ridership, i.e., Origin-Destination (OD) one and Destination-Origin (DO) one. However, complete OD matrices of previous time intervals can not be obtained immediately in online metro systems, and conventional methods only used limited information to forecast the future OD and DO ridership separately. In this work, we proposed a novel neural network module termed Heterogeneous Information Aggregation Machine (HIAM), which fully exploits heterogeneous information of historical data (e.g., incomplete OD matrices, unfinished order vectors, and DO matrices) to jointly learn the evolutionary patterns of OD and DO ridership. Specifically, an OD modeling branch estimates the potential destinations of unfinished orders explicitly to complement the information of incomplete OD matrices, while a DO modeling branch takes DO matrices as input to capture the spatial-temporal distribution of DO ridership. Moreover, a Dual Information Transformer is introduced to propagate the mutual information among OD features and DO features for modeling the OD-DO causality and correlation. Based on the proposed HIAM, we develop a unified Seq2Seq network to forecast the future OD and DO ridership simultaneously. Extensive experiments conducted on two large-scale benchmarks demonstrate the effectiveness of our method for online metro origin-destination prediction.

show abstract

Section: Traffic Origin-destination Predictionmentioning

confidence: 99%

Section: ) Transaction Datamentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Online Metro Origin-Destination Prediction via Heterogeneous Information Aggregation

Liu¹,

Zhu²,

Li³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Later, a multi-view learning framework was developed in [18] for taxi demand prediction, which incorporates CNNs and RNNs. Noursalehi et al [19] introduced a spatiotemporal model for the origin-destination (OD) demand prediction of urban rail systems with convolutional layers to capture the spatial dependencies within OD matrices. While CNNs work well for correlations in Euclidean space (e.g., a spatial grid), they are not applicable to non-Euclidean space, such as irregular service zones and unevenly distributed transit stations.…”

Section: Single-mode Demand Predictionmentioning

confidence: 99%

Joint Demand Prediction for Multimodal Systems: A Multi-task Multi-relational Spatiotemporal Graph Neural Network Approach

Liang,

Huang,

Zhao

2021

Preprint

View full text Add to dashboard Cite

Dynamic demand prediction is crucial for the efficient operation and management of urban transportation systems. Extensive research has been conducted on single-mode demand prediction, ignoring the fact that the demands for different transportation modes can be correlated with each other. Despite some recent efforts, existing approaches to multimodal demand prediction are generally not flexible enough to account for multiplex networks with diverse spatial units and heterogeneous spatiotemporal correlations across different modes. To tackle these issues, this study proposes a multi-relational spatiotemporal graph neural network (ST-MRGNN) for multimodal demand prediction. Specifically, the spatial dependencies across modes are encoded with multiple intra-and inter-modal relation graphs. A multi-relational graph neural network (MRGNN) is introduced to capture cross-mode heterogeneous spatial dependencies, consisting of generalized graph convolution networks to learn the message passing mechanisms within relation graphs and an attention-based aggregation module to summarize different relations. We further integrate MRGNNs with temporal gated convolution layers to jointly model heterogeneous spatiotemporal correlations. Extensive experiments are conducted using real-world subway and ride-hailing datasets from New York City, and the results verify the improved performance of our proposed approach over existing methods across modes. The improvement is particularly large for demand-sparse locations. Further analysis of the attention mechanisms of ST-MRGNN also demonstrates its good interpretability for understanding cross-mode interactions.

show abstract

FMSYS: Fine-Grained Passenger Flow Monitoring in a Large-Scale Metro System Based on AFC Smart Card Data

Sun,

Zhao,

Zhang

et al. 2024

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Dynamic Origin-Destination Prediction in Urban Rail Systems: A Multi-Resolution Spatio-Temporal Deep Learning Approach

Cited by 62 publications

References 45 publications

Online Metro Origin-Destination Prediction via Heterogeneous Information Aggregation

Online Metro Origin-Destination Prediction via Heterogeneous Information Aggregation

Joint Demand Prediction for Multimodal Systems: A Multi-task Multi-relational Spatiotemporal Graph Neural Network Approach

FMSYS: Fine-Grained Passenger Flow Monitoring in a Large-Scale Metro System Based on AFC Smart Card Data

Contact Info

Product

Resources

About