融合时空特征的多模态车辆轨迹预测方法

doi:10.3778/j.issn.1002-8331.2311-0240

摘要/Abstract

摘要： 针对考虑车辆行驶不确定性的轨迹分布准确快速预测问题，提出了一种融合时空特征的多模态车辆轨迹预测方法（GCNTA）。利用空间关联度系数和图卷积神经网络（GCN）实现空间关联特征提取。构建具有时间注意力机制的时域卷积网络（TCN）完成时间特征提取。通过特征融合门控单元实现每个时间步长对应时空特征的自适应融合，并利用门控循环单元（GRU）网络构建解码器进一步生成未来车辆轨迹的概率分布。利用公开的NGSIM数据集对所提出模型进行消融实验及预测精度分析。仿真结果表明，GCNTA模型在预测误差均方根（RMSE）平均值相比GCN、图注意力网络（GAT）和长短期记忆网络（LSTM）模型分别减少15.6%、16.3%和23.8%。

关键词: 车辆轨迹预测, 深度学习, 图神经网络, 时域卷积网络, 注意力机制

Abstract: Regarding the issue of accurate and rapid prediction of trajectory distributions taking into account the uncertainty of vehicle movement, a multimodal trajectory prediction method fusing spatio-temporal features extracted by graph convolutional networks and temporal attentions (GCNTA) is proposed. Firstly, the spatial correlation coefficients and the graph convolutional neural network (GCN) are utilized to extract the spatial correlation features. Secondly, the temporal convolutional network (TCN) network with temporal attention mechanism is constructed to extract the temporal features. Then, the adaptive fusion of spatio-temporal features at each time step is realized by the feature fusion gating unit, and a decoder is constructed to further generate the probability distribution of future vehicle trajectories by using gated recurrent unit (GRU) network. Finally, the proposed model is analyzed based on the ablation experiments and the prediction accuracy by using the publicly available NGSIM dataset. The simulation results show that the GCNTA model can reduce 15.6%, 16.3%, and 23.8% in the root mean square error (RMSE) average compared to the GCN, graph attention network (GAT), and long short-term memory network (LSTM) models, respectively.

Key words: vehicle trajectory prediction, deep learning, graph neural network, temporal convolutional network, attention mechanism

史昕, 王浩泽, 纪艺, 马峻岩. 融合时空特征的多模态车辆轨迹预测方法[J]. 计算机工程与应用, 2025, 61(7): 325-333.

SHI Xin, WANG Haoze, JI Yi, MA Junyan. Multimodal Vehicle Trajectory Prediction Method with Fusion of Spatio-Temporal Features[J]. Computer Engineering and Applications, 2025, 61(7): 325-333.

参考文献

[1] 吴翊恺, 胡启洲, 吴啸宇. 车联网背景下的机动车辆轨迹预测模型[J]. 东南大学学报(自然科学版), 2022, 52(6): 1199-1208.
WU Y K, HU Q Z, WU X Y. Vehicle trajectory prediction model in the context of Internet of vehicles[J]. Journal of Southeast University (Natural Science Edition), 2022, 52(6): 1199-1208.
[2] 李雪松, 张锲石, 宋呈群, 等. 自动驾驶场景下的轨迹预测技术综述[J]. 计算机工程, 2023, 49(5): 1-11.
LI X S, ZHANG Q S, SONG C Q, et al. Review of trajectory prediction technology in autonomous driving scenes[J]. Computer Engineering, 2023, 49(5): 1-11.
[3] LIN C F, ULSOY A G, LEBLANC D J. Vehicle dynamics and external disturbance estimation for vehicle path prediction[J]. IEEE Transactions on Control Systems Technology, 2000, 8(3): 508-518.
[4] 乔少杰, 韩楠, 朱新文, 等. 基于卡尔曼滤波的动态轨迹预测算法[J]. 电子学报, 2018, 46(2): 418-423.
QIAO S J, HAN N, ZHU X W, et al. A dynamic trajectory prediction algorithm based on Kalman filter[J]. Acta Electronica Sinica, 2018, 46(2): 418-423.
[5] 高建, 毛莺池, 李志涛. 基于高斯混合-时间序列模型的轨迹预测[J]. 计算机应用, 2019, 39(8): 2261-2270.
GAO J, MAO Y C, LI Z T. Trajectory prediction based on Gauss mixture time series model[J]. Journal of Computer Applications, 2019, 39(8): 2261-2270.
[6] 季学武, 费聪, 何祥坤, 等. 基于LSTM网络的驾驶意图识别及车辆轨迹预测[J]. 中国公路学报, 2019, 32(6): 34-42.
JI X W, FEI C, HE X K, et al. Intention recognition and trajectory prediction for vehicles using LSTM network[J]. China Journal of Highway and Transport, 2019, 32(6): 34-42.
[7] 吴晓建, 危一华, 王爱春, 等. 基于融合Dropout与注意力机制的LSTM-GRU车辆轨迹预测[J]. 湖南大学学报(自然科学版), 2023, 50(4): 65-75.
WU X J, WEI Y H, WANG A C, et al. Vehicle trajectory prediction based on LSTM-GRU integrating Dropout and attention mechanism[J]. Journal of Hunan University (Natural Sciences), 2023, 50(4): 65-75.
[8] 方华珍, 刘立, 肖小凤, 等. 混合示教长短时记忆网络的车辆轨迹预测研究[J]. 交通运输系统工程与信息, 2023, 23(4): 80-87.
FANG H Z, LIU L, XIAO X F, et al. Vehicle trajectory prediction based on mixed teaching force long short-term memory[J]. Journal of Transportation Systems Engineering and Information Technology, 2023, 23(4): 80-87.
[9] 黄玲, 崔躜, 游峰, 等. 适用于多车交互场景的车辆轨迹预测模型[J]. 吉林大学学报(工学版), 2024, 54(5): 1188-1195.
HUANG L, CUI Z, YOU F, et al. Vehicle trajectory prediction model for multi-vehicle interaction scenario[J]. Journal of Jilin University (Engineering and Technology Edition), 2024, 54(5): 1188-1195.
[10] 李文礼, 韩迪, 石晓辉, 等. 基于时-空注意力机制的车辆轨迹预测[J]. 中国公路学报, 2023, 36(1): 226-239.
LI W L, HAN D, SHI X H, et al. Vehicle trajectory prediction based on spatial-temporal attention mechanism[J]. China Journal of Highway and Transport, 2023, 36(1): 226-239.
[11] DIEHL F, THOMAS B, MICHAEL T L, et al. Graph neural networks for modelling traffic participant interaction[C]// Proceedings of the Intelligent Vehicles Symposium (IV), 2019: 695-701.
[12] MO X Y, XING Y, LV C. Graph and recurrent neural network-based vehicle trajectory prediction for highway driving[J]. arXiv:2107.03663, 2021.
[13] LI X, YING X W, CHUAH M C. GRIP: graph-based interaction-aware trajectory prediction[C]//Proceedings of the 2019 IEEE Intelligent Transportation Systems Conference. Piscataway: IEEE, 2019: 3960-3966.
[14] ZHAO T Y, XU Y F, MONFORT M, et al. Multi-agent tensor fusion for contextual trajectory prediction[C]//Proceedings of the 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2019: 12118-12126.
[15] 陈泽, 杨铀. 多重时空交互下的行人轨迹预测[J]. 华中科技大学学报(自然科学版), 2023, 51(9): 61-67.
CHEN Z, YANG Y. Pedestrian trajectory prediction based on multiple spatio-temporal interaction[J]. Journal of Huazhong University of Science and Technology (Natural Science Edition), 2023, 51(9): 61-67.
[16] 陈晓伟, 李煊鹏, 张为公. 基于动态图注意力的车辆轨迹预测研究[J]. 汽车技术, 2024(3): 24-30.
CHEN X W, LI X P, ZHANG W G. Research on vehicle trajectory prediction based on dynamic graph attention[J]. Automobile Technology, 2024(3): 24-30.
[17] 曾伟良, 陈漪皓, 姚若愚, 等. 时空图注意力网络在交叉口车辆轨迹预测的应用[J]. 计算机科学, 2021, 48(S1): 334-341.
ZENG W L, CHEN Y H, YAO R Y, et al. Application of spatial-temporal graph attention networks in trajectory prediction for vehicles at intersections[J]. Computer Science, 2021, 48(S1): 334-341.
[18] 刘建敏, 林晖, 汪晓丁. 基于图注意力机制的无地图场景轨迹预测方法[J]. 计算机工程, 2024, 50(7): 144-153.
LIU J M, LIN H, WANG X D. Graph attention mechanism-based method for trajectory prediction in map-free scenes[J]. Computer Engineering, 2024, 50(7): 144-153.
[19] SHENG Z H, XU Y W, XUE S B, et al. Graph-based spatial-temporal convolutional network for vehicle trajectory prediction in autonomous driving[J]. IEEE Transactions on Intelligent Transportation Systems, 2022, 23(10): 17654-17665.
[20] WU Z H, PAN S R, CHEN F W, et al. A comprehensive survey on graph neural networks[J]. IEEE Transactions on Neural Networks and Learning Systems, 2021, 32(1): 4-24.
[21] LIU M P, QIN H, CAO R, et al. Short-term load forecasting based on improved TCN and DenseNet[J]. IEEE Access, 2022, 10: 115945-115957.
[22] MOHAMED A, QIAN K, ELHOSEINY M, et al. Social-STGCNN: a social spatio-temporal graph convolutional neural network for human trajectory prediction[C]//Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2020: 14412-14420.
[23] DEO N, TRIVEDI M M. Multi-modal trajectory prediction of surrounding vehicles with maneuver based LSTMs[C]//Proceedings of the 2018 IEEE Intelligent Vehicles Symposium. Piscataway: IEEE, 2018: 1179-1184.
[24] DEO N, TRIVEDI M M. Convolutional social pooling for vehicle trajectory prediction[C]//Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops. Piscataway: IEEE, 2018: 1468-1476.