Pedestrian trajectory prediction based on multiple spatio-temporal interaction

被引：0

作者：

Chen Z. ^{[1
]}

Yang Y. ^{[1
]}

机构：

[1] School of Electronic Information and Communications, Huazhong University of Science and Technology, Wuhan

来源：

Huazhong Keji Daxue Xuebao (Ziran Kexue Ban)/Journal of Huazhong University of Science and Technology (Natural Science Edition) | 2023年 / 51卷 / 09期

关键词：

attention mechanism; deep learning; graph construction; graph neural network; trajectory prediction;

D O I：

10.13245/j.hust.230902

中图分类号：

学科分类号：

摘要：

The future trajectory of pedestrians is continuously affected by the surrounding pedestrians，and this kind of contribution hardly be involved in the consideration．In order to improve the prediction performance，a pedestrian trajectory prediction method was proposed based on a multi-spatial graph neural network．The proposed approach included a multiple spatial fusion module that integrates different spatio-temporal information together to construct a multi-spatial graph neural network．The module considered the influence of the past trajectories of other pedestrians to obtain interaction information that spanned both time and space between pedestrians．Subsequently，transformer was applied to the multi-spatial graph network to adaptively measure the interaction between pedestrians across time and space．The proposed model was compared with state-of-the-arts on the ETH and UCY datasets．The experimental results show that the proposed model is 12% and 15% more prior than state-of-the-arts on metrics of ADE (average displacement error) and FDE (final displacement error)，and can generate consistent trajectories for pedestrians. © 2023 Huazhong University of Science and Technology. All rights reserved.

引用

下载

页码：61 / 67

页数：6

共 16 条

[1] BAI H, CAI S, YE N, Intention-aware online POMDP planning for autonomous driving in a crowd[C], Proc of IEEE International Conference on Robotics and Automation, pp. 454-460, (2015)
[2] KRUSE T, PANDEY A K, ALAMI R, Humanaware robot navigation：a survey[J], Robotics and Autonomous Systems, 61, 12, pp. 1726-1743, (2013)
[3] YASUNO M, RYOUSUKE S, YASUDA N, Pedestrian detection and tracking in far infrared images[C], Proc of International Conference on Intelligent Transportation Systems, pp. 131-136, (2005)
[4] ALAHI A, GOEL K, RAMANATHAN V, Social LSTM：human trajectory prediction in crowded spaces[C], Proc of IEEE Conference on Computer Vision and Pattern Recognition, pp. 961-971, (2016)
[5] GUPTA A，, JOHNSON J，, FEI-FEI L， et al． Social GAN：socially acceptable trajectories with generative adversarial networks[C], Proc of IEEE Conference on Computer Vision and Pattern Recognition, pp. 2255-2264, (2018)
[6] ZHANG P，, OUYANG W，, ZHANG P, Sr-LSTM： state refinement for LSTM towards pedestrian trajectory prediction[C], Proc of IEEE Conference on Computer Vision and Pattern Recognition, pp. 12077-12086, (2019)
[7] HUANG Y，, BI H，, LI Z, STGAT： modeling spatial-temporal interactions for human trajectory prediction[C], Proc of IEEE International Conference on Computer Vision, pp. 6271-6280, (2019)
[8] MOHAMED A, QIAN K, ELHOSEINY M, Social-STGCNN： a social spatio-temporal graph convolutional neural network for human trajectory prediction[C], Proc of IEEE Conference on Computer Vision and Pattern Recognition, pp. 14412-14420, (2020)
[9] LEA C，, FLYNN M D，, VIDAL R, Temporal convolutional networks for action segmentation and detection[C], Proc of IEEE Conference on Computer Vision and Pattern Recognition, pp. 1003-1012, (2017)
[10] WANG C, CAI S, TAN G．, GraphTCN：spatio-temporal interaction modeling for human trajectory prediction [C], Proc of IEEE Winter Conference on Applications of Computer Vision, pp. 3450-3459, (2021)

← 1 2 →