TY - JOUR
T1 - SPRNN
T2 - A spatial–temporal recurrent neural network for crowd flow prediction
AU - Tang, Gaozhong
AU - Li, Bo
AU - Dai, Hong Ning
AU - Zheng, Xi
N1 - Funding Information:
This research was partially supported by the National Natural Science Foundation of China (11627802, 11872185 and 12172134), by the State Scholarship Fund of China Scholarship Council (201806155 022), by Macao Science and Technology Development Fund under Macao Funding Scheme for Key R & D Projects (0025/2019/AKP), by an Australian Research Council (ARC) Discovery Project (DP210102447), by an ARC Linkage Project (LP190100676), and by a DATA61 project (Data61 CRP C020996).
Publisher Copyright:
© 2022 Elsevier Inc.
PY - 2022/10
Y1 - 2022/10
N2 - The capability of predicting the future trends of crowds has rendered crowd flow prediction more critical in building intelligent transportation systems, and attracted substantial research efforts. The trend of crowd flows is closely related to time and the urban topography. Therefore, extracting and leveraging both spatial features and temporal features are key gradients for effectively predicting crowd flows. Many previous works extract spatial features from crowd-flow data in an iteration way. As a result, models suffer from a heavy computation cost while ignoring details of road topology and structure information. Meanwhile, temporal features, including short-term features and long-term features, are separately extracted. The fusion of all features at the last stage before accomplishing the prediction also neglects the underlying associativity between various features. To address the limitations, we leverage spatial features by extracting structural information of road structures, such as road connection, road density, road width, etc. Rather than extracting spatial features from crowd-flow data, we capture them from images of city maps by adopting convolutional neural networks. Moreover, we implement a new sequence feature fusion mechanism to merge both spatial features and temporal features from various time scales so as to predict crowd flows. We conduct extensive experiments to evaluate our model on three benchmark datasets. The experimental results demonstrate that the model outperforms 15 state-of-the-art methods. The source code is available at: https://github.com/CVisionProcessing/SPRNN.
AB - The capability of predicting the future trends of crowds has rendered crowd flow prediction more critical in building intelligent transportation systems, and attracted substantial research efforts. The trend of crowd flows is closely related to time and the urban topography. Therefore, extracting and leveraging both spatial features and temporal features are key gradients for effectively predicting crowd flows. Many previous works extract spatial features from crowd-flow data in an iteration way. As a result, models suffer from a heavy computation cost while ignoring details of road topology and structure information. Meanwhile, temporal features, including short-term features and long-term features, are separately extracted. The fusion of all features at the last stage before accomplishing the prediction also neglects the underlying associativity between various features. To address the limitations, we leverage spatial features by extracting structural information of road structures, such as road connection, road density, road width, etc. Rather than extracting spatial features from crowd-flow data, we capture them from images of city maps by adopting convolutional neural networks. Moreover, we implement a new sequence feature fusion mechanism to merge both spatial features and temporal features from various time scales so as to predict crowd flows. We conduct extensive experiments to evaluate our model on three benchmark datasets. The experimental results demonstrate that the model outperforms 15 state-of-the-art methods. The source code is available at: https://github.com/CVisionProcessing/SPRNN.
KW - Crowd flow prediction
KW - Gated recurrent unit
KW - Road structural information
KW - Spatial feature
KW - Temporal feature
UR - http://www.scopus.com/inward/record.url?scp=85139225309&partnerID=8YFLogxK
U2 - 10.1016/j.ins.2022.09.053
DO - 10.1016/j.ins.2022.09.053
M3 - Journal article
AN - SCOPUS:85139225309
SN - 0020-0255
VL - 614
SP - 19
EP - 34
JO - Information Sciences
JF - Information Sciences
ER -