联合TCN和时空多头注意机制的车辆轨迹预测模型

doi:10.3778/j.issn.1002-8331.2405-0015

摘要/Abstract

摘要： 车辆轨迹预测为自动驾驶系统提供决策和规划的基础数据，它是自动驾驶过程的一个重要环节。然而，车辆轨迹预测过程存在复杂的空间交互性和时间相关性，给轨迹预测带来了巨大挑战。因此，提出了一种基于时空多头注意机制和时间卷积网络（temporal convolutional network，TCN）的车辆轨迹预测模型。将门控机制和TCN结合，经过多头注意机制后进行堆叠以提取不同层次的时间特征，并分配相应的权重。将相邻车辆的历史轨迹以栅格图的形式进行卷积操作和coordinate attention（CA）操作以提取空间交互特征。预测未来不同机动意图的概率，并将其和提取到的时空特征输入到基于长短期记忆网络（long short-term memory network，LSTM）的解码器中获得未来轨迹。所提模型在下一代仿真（next generation simulation，NGSIM）数据集进行了实验评估。与4种相似模型相比，提出的模型预测误差最高降低了17.8%。

关键词: 车辆轨迹预测, 自动驾驶, 时空多头注意机制, 时空特征

Abstract: Vehicle trajectory prediction provides basic data for decision-making and planning of the automatic driving system, which is an important part of the automatic driving process. However, the process of vehicle trajectory prediction has complex spatial interaction and temporal correlation, which brings great challenges to trajectory prediction. Therefore, a vehicle trajectory prediction model based on temporal convolutional network (TCN) and multi-head attention mechanism is proposed. Firstly, the gating mechanism and TCN are combined, and stacked to extract temporal features of different levels and assign corresponding weights after multi-head attention mechanism. Then, the historical tracks of adjacent vehicles are convolved and coordinate attention (CA) operations are carried out in the form of grid graphs to extract the spatial interaction features. Finally, the probability of different maneuvering intentions in the future is predicted, and the extracted spatial-temporal features are input into the decoder based on the long short-term memory network (LSTM) to obtain the future trajectory. The proposed model is evaluated experimentally on the next generation simulation (NGSIM) dataset. Compared to other four similar models, the prediction error of the proposed model is reduced by up to 17.8%.

Key words: vehicle trajectory prediction, automatic driving, spatial-temporal multiple attention mechanism, spatio-temporal features

宋绍剑, 徐佳敏, 李刚, 李国进. 联合TCN和时空多头注意机制的车辆轨迹预测模型[J]. 计算机工程与应用, 2025, 61(12): 352-358.

SONG Shaojian, XU Jiamin, LI Gang, LI Guojin. Vehicle Trajectory Prediction Model with TCN and Spatial-Temporal Multi-Head Attention Mechanism[J]. Computer Engineering and Applications, 2025, 61(12): 352-358.

参考文献

[1] 国家统计局. 中国统计年鉴. 2011 (总第30期)[M]. 北京: 中国统计出版社, 2011.
National Bureau of Statistics. Chinese statistical yearbook. 2011 (Total No. 30) [M]. Beijing: China Statistics Press, 2011.
[2] 孟宪伟, 唐进君, 王喆. 考虑换道意图的LSTM-AdaBoost车辆轨迹预测模型[J]. 计算机工程与应用, 2022, 58(13): 280-287.
MENG X W, TANG J J, WANG Z. Trajectory prediction of vehicles based on LSTM-AdaBoost model considering lane-changing intention[J]. Computer Engineering and Applications, 2022, 58(13): 280-287.
[3] MESSAOUD K, YAHIAOUI I, VERROUST-BLONDET A, et al. Attention based vehicle trajectory prediction[J]. IEEE Transactions on Intelligent Vehicles, 2021, 6(1): 175-185.
[4] HUANG Y J, DU J T, YANG Z R, et al. A survey on trajectory-prediction methods for autonomous driving[J]. IEEE Transactions on Intelligent Vehicles, 2022, 7(3): 652-674.
[5] ZHANG K P, ZHAO L, DONG C X, et al. AI-TP: attention-based interaction-aware trajectory prediction for autonomous driving[J]. IEEE Transactions on Intelligent Vehicles, 2023, 8(1): 73-83.
[6] LI Z R, LU C, YI Y T, et al. A hierarchical framework for interactive behaviour prediction of heterogeneous traffic participants based on graph neural network[J]. IEEE Transactions on Intelligent Transportation Systems, 2022, 23(7): 9102-9114.
[7] GUO Y H, KALIDINDI V V, ARIEF M, et al. Modeling multi-vehicle interaction scenarios using Gaussian random field[C]//Proceedings of the 2019 IEEE Intelligent Transportation Systems Conference. Piscataway: IEEE, 2019: 3974-3980.
[8] MOZAFFARI S, AL-JARRAH O Y, DIANATI M, et al. Deep learning-based vehicle behavior prediction for autonomous driving applications: a review[J]. IEEE Transactions on Intelligent Transportation Systems, 2022, 23(1): 33-47.
[9] ZYNER A, WORRALL S, NEBOT E. A recurrent neural network solution for predicting driver intention at unsignalized intersections[J]. IEEE Robotics and Automation Letters, 2018, 3(3): 1759-1764.
[10] LIN L, LI W Z, BI H K, et al. Vehicle trajectory prediction using LSTMs with spatial-temporal attention mechanisms[J]. IEEE Intelligent Transportation Systems Magazine, 2022, 14(2): 197-208.
[11] XUE Q F, LI S Y, LI X P, et al. Hierarchical motion encoder-decoder network for trajectory forecasting[J]. arXiv:2111. 13324, 2021.
[12] GUPTA A, JOHNSON J, LI F F, et al. Social GAN: socially acceptable trajectories with generative adversarial networks[C]//Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2018: 2255-2264.
[13] DING W C, CHEN J, SHEN S J. Predicting vehicle behaviors over an extended horizon using behavior interaction network[C]//Proceedings of the 2019 International Conference on Robotics and Automation. Piscataway: IEEE, 2019: 8634-8640.
[14] MESSAOUD K, DEO N, TRIVEDI M M, et al. Trajectory prediction for autonomous driving based on multi-head attention with joint agent-map representation[C]//Proceedings of the 2021 IEEE Intelligent Vehicles Symposium. Piscataway: IEEE, 2021: 165-170.
[15] LIU Y C, ZHANG J H, FANG L J, et al. Multimodal motion prediction with stacked transformers[C]//Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2021: 7573-7582.
[16] NIKHIL N, MORRIS B T. Convolutional neural network for trajectory prediction[J]. arXiv:1809.00696, 2018.
[17] STROHBECK J, BELAGIANNIS V, MüLLER J, et al. Multiple trajectory prediction with deep temporal and spatial convolutional neural networks[C]//Proceedings of the 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway: IEEE, 2020: 1992-1998.
[18] 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: 1549-15498.
[19] 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.
[20] HUANG Y F, BI H K, LI Z X, et al. STGAT: modeling spatial-temporal interactions for human trajectory prediction[C]//Proceedings of the 2019 IEEE/CVF International Conference on Computer Vision. Piscataway: IEEE, 2019: 6271-6280.
[21] VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[J]. arXiv:1706.03762, 2017.
[22] HU Y, CHEN X B. Intention-aware transformer with adaptive social and temporal learning for vehicle trajectory prediction[C]//Proceedings of the 2022 26th International Conference on Pattern Recognition. Piscataway: IEEE, 2022: 3721-3727.