摘 要:针对传统的轨迹预测存在精度低和计算较复杂的问题,文章提出一种基于自动编码器(AE)和门控循环单元(GRU)模型的数据驱动方法,利用历史信息和各种轨迹属性预测轨迹位置。该方法将数据预处理层、AE 层和 GRU 层与定制的批处理过程融合在一起。该模型在真实轨迹数据集上训练。通过与现有预测方法进行比较,结果验证所提模型性能相比于RNN,AE-RNN,LSTM 和 GRU 有显著的提高。
关键词:轨迹预测;自动编码器;门控循环单元;数据驱动
DOI:10.19850/j.cnki.2096-4706.2021.02.036
中图分类号:TP183 文献标识码:A 文章编号:2096-4706(2021)02-0149-05
Research on Trajectory Prediction Based on AE and GRU Network
ZHANG Hao,LIU Daming
(School of Computer Science and Technology,Shanghai University of Electric Power,Shanghai 201303,China)
Abstract:Aiming at the problems of low precision and computational complexity in traditional trajectory prediction,a data driven method based on auto encoder(AE)and gated recurrent unit(GRU)model was proposed,which used historical information and various trajectory attributes to predict trajectory location. This method fuses the data preprocessing layer,AE layer and GRU layer with the custom batch process. The model is trained on real trajectory data sets. Compared with the existing prediction methods,the results show that the performance of the proposed model is significantly better than that of RNN,AE-RNN,LSTM and GRU.
Keywords:trajectory prediction;AE;GRU;data driven
参考文献:
[1] ALAHI A,GOEL K,RAMANATHAN V,et al. Social LSTM:Human Trajectory Prediction in Crowded Spaces [C]//2016IEEE Conference on Computer Vision and Pattern Recognition(CVPR).Las Vegas:IEEE,2016:961-971.
[2] 尹博南,艾元,彭木根 . 雾无线接入网:架构、原理和挑战 [J]. 电信科学,2016,32(6):20-27.
[3] PENG M G,LI Y,JIANG J M,et al. Heterogeneous cloud radio access networks:a new perspective for enhancing spectral and energy efficiencies [J].IEEE Wireless Communications,2014,21(6):126-135.
[4] ZHANG Z F,ZHANG P R,LIU D,et al. SRSM-Based Adaptive Relay Selection for D2D Communications [J]. IEEE Internet of Things Journal,2017,5(4):2323-2332.
[5] ZHANG F S,JIN B H,WANG Z Y,et al. On Geocasting over Urban Bus-Based Networks by Mining Trajectories [J].IEEE Transactions on Intelligent Transportation Systems,2016,17(6):1734–1747.
[6] SCHWARTING W,ALONSO-MORA J,RUS D. Planning and Decision-Making for Autonomous Vehicles [J] Annual Review of Control,Robotics,and Autonomous Systems,2018(1):187–210.
[7] ZHU L,XU C Q,GUAN J F,et al. Finding top-k similar users based on Trajectory-Pattern model for personalized service recommendation [C]//2016 IEEE International Conference on Communications Workshops (ICC).Kuala Lumpur:IEEE,2016:553–558.
[8] ZHENG Y. Trajectory data mining:an overview [J].ACM Transactions on Intelligent Systems and Technology,2015,6(3):1-41.
[9] HOUENOU A,BONNIFAIT P,CHERFAOUI V,et al.Vehicle trajectory prediction based on motion model and maneuver recognition [C]//2013 IEEE/RSJ International Conference on Intelligent Robots and Systems.Tokyo:IEEE,2013:4363–4369.
[10] GIACOMETTI A,SOULET A. Frequent Pattern Outlier Detection Without Exhaustive Mining [C]//Advances in Knowledge Discovery and Data Mining.Berlin:Springer-Verlag,2016:196–207.
[11] LI Y,LIU J,CAO B,et al. Joint Optimization of Radio and Virtual Machine Resources With Uncertain User Demands in Mobile Cloud Computing [J].IEEE Transactions on Multimedia,2018,20(9):2427–2438.
[12] PENG M G,ZHANG K C,JIANG J M,et al. EnergyEfficient Resource Assignment and Power Allocation in Heterogeneous Cloud Radio Access Networks [J].IEEE Transactions on Vehicular Technology,2015,64(11):5275–5287.
[13] CAO B,XIA S C,HAN J W,et al. A Distributed Game Methodology for Crowdsensing in Uncertain Wireless Scenario [J].IEEE Transactions on Mobile Computing.2020,19(1):15–28.
[14] CAO B,ZHANG L,LI Y,et al. Intelligent Offloading in Multi-Access Edge Computing:A State-of-the-Art Review and Framework [J].IEEE Communications Magazine.2019,57(3):56–62.
[15] ZHANG Z F,ZOU Y,GAN C Q. Textual sentiment analysis via three different attention convolutional neural networks and cross-modality consistent regression [J].Neurocomputing.2018,275:1407–1415.
[16] LIU T,LI Z R,YU C X,et al. NIRS feature extraction based on deep auto-encoder neural network [J].Infrared Physics & Technology,2017,87::124-128.
[17] SHAO H D,JIANG H K,ZHAO H W,et al. A novel deep autoencoder feature learning method for rotating machinery fault diagnosis [J].Mechanical Systems and Signal Processing,2017,95:187-204.
[18] MENG Q X,CATCHPOOLE D,SKILLICOM D,et al. Relational autoencoder for feature extraction [C]//2017 International Joint Conference on Neural Networks(IJCNN).Anchorage:IEEE,2017:364–371.
[19] HIDASI B,KARATZOGLOU A,BALTRUNAS L,et al. Session-based Recommendations with Recurrent Neural Networks [J/OL].arXiv:1511.06939 [cs.LG].(2016-03-29).https://arxiv.org/ abs/1511.06939.
作者简介:张浩(1996—),男,汉族,安徽蚌埠人,硕士研究生在读,研究方向:轨迹预测,物联网技术;刘大明(1971—),男,汉族,上海人,副教授,博士,研究方向:物联网技术,嵌入式系统与设计,智能工业机器人等。