-
摘要: 为从路网速度中完整提取路段速度的时空特征,实现高精度路段速度预测,通过调用在线地图的路径规划应用程序接口,采集路段的在线地图速度;利用图卷积神经网络(GCN)提取空间特征,利用长短期记忆(LSTM)神经网络提取时间特征,建立面向在线地图的GCN-LSTM神经网络,提取路段速度的时空特征,预测路段速度;为测试面向在线地图的GCN-LSTM神经网络表现,并评价在线地图下GCN-LSTM神经网络的优势与面向检测器速度预测模型的可替代性,以局部路网为例分析模型表现,并对比在线地图下不同模型的表现与不同数据源下近似模型的表现。研究结果表明:GCN-LSTM神经网络在训练集和测试集上的平均绝对误差(MAE)均低于5,均方根误差(RMSE)均低于6,平均绝对百分比误差(MAPE)均低于30%,训练误差和测试误差均处于较低水平,总体表现良好;GCN-LSTM神经网络的路段MAPE服从Gumbel分布,均值均落在19%±4%之间,85%分位点均落在34%±5%之间,2项指标均处于较低水平,个体表现良好;在面向在线地图的速度预测模型中,GCN-LSTM神经网络的MAE、RMSE、MAPE以及MAPE拟合曲线均值、85%分位点最低,总体和个体表现均为最佳,在面向在线地图的速度预测中具有一定优势;在近似模型中,GCN-LSTM神经网络的MAE、RMSE、MAPE以及MAPE拟合曲线均值、85%分位点最低,总体和个体表现均为最佳,则面向在线地图速度预测的可靠性高,可代替面向检测器的速度预测。Abstract: Online map speeds of roads were collected by calling the path-planning application programming interface of the online map to completely extract the spatio-temporal features of the road speed from road network speed and then achieve high-precision road speed prediction. The spatial features were extracted using a graph convolutional network (GCN), and the temporal features were extracted using a long short-term memory (LSTM) neural network. An online map-based GCN-LSTM neural network was established, the spatio-temporal features of the road speed were extracted, and the road speed was predicted. The performance of the online map-based GCN-LSTM neural network was assessed, and the advantages of the online map-based GCN-LSTM neural network and the substitutability of the detector-based speed prediction model were evaluated. By using the local road network as an example, the performance of the model was analyzed, and the performances of different online map-based models and similar models with different data sources were compared. Analysis results show that the mean absolute errors(MAEs) of the GCN-LSTM neural network are lower than 5, the root mean square errors (RMSEs) are lower than 6, and the mean absolute percentage errors (MAPEs) are lower than 30% in the training and testing sets. Hence, the training and testing errors are low, indicating good comprehensive performance. The MAPE of the GCN-LSTM neural network of the roads follows a Gumbel distribution, whose mean ranges between 19%±4%, and the 85% quantile ranges between 34%±5%. Hence, both indexes are low, indicating good individual performance. Among the online map-based speed prediction models, the MAE, RMSE, MAPE, mean, and 85% quantile of the MAPE fitting curve of the GCN-LSTM neural network have the lowest values. Hence, its comprehensive and individual performances are the best, and it exhibits advantages in online map-based speed prediction. Among the similar models, the MAE, RMSE, MAPE, mean, and 85% quantile of the MAPE fitting curve of the GCN-LSTM neural network have the lowest values. Hence, its comprehensive and individual performances are the best. Furthermore, the reliability of online map-based speed prediction is high, so that it can be used as a substitute for detector-based speed prediction. 4 tabs, 13 figs, 30 refs.
-
表 1 百度地图速度
Table 1. Speed of Baidu map
(km·h-1) 时刻 17:00 17:05 … 19:30 路段1速度 26.0 25.2 … 25.8 路段2速度 28.0 27.1 … 27.1 ⋮ ⋮ ⋮ ⋮ ⋮ 表 2 Gumbel曲线拟合结果
Table 2. Results of Gumbel curve fitting
数据集 拟合结果 17:30 18:00 18:30 19:00 训练集 R2 0.83 0.80 0.66 0.77 μ/% 19.41 22.27 21.23 19.91 85%分位点/% 32.82 34.50 36.32 31.00 测试集 R2 0.71 0.74 0.89 0.63 μ/% 18.53 21.33 17.87 16.26 85%分位点/% 32.36 38.52 29.44 31.38 表 3 在线地图下不同模型的总体表现对比
Table 3. Comprehensive performance comparison of different online map-based models
模型名称 MAE RMSE MAPE/% 17:30 18:00 18:30 19:00 17:30 18:00 18:30 19:00 17:30 18:00 18:30 19:00 SVM 6.48 6.20 5.01 4.97 8.05 7.66 6.44 6.29 35.98 37.55 30.97 33.82 LSTM 5.60 5.58 5.28 6.34 6.89 6.92 6.48 7.43 27.01 32.89 28.62 37.31 GCN 6.73 6.51 5.41 6.87 8.03 8.22 6.57 8.02 30.64 37.94 29.41 37.44 GCN-Seq2Seq 5.71 5.34 4.77 5.75 6.94 6.64 6.00 6.85 27.29 31.95 26.86 34.35 GCN-LSTM 4.17 4.44 3.99 4.28 5.33 5.61 5.13 5.27 22.48 27.69 24.28 28.44 表 4 不同数据源下近似模型的总体表现对比
Table 4. Comprehensive performance comparison of similar models with different data sources
模型名称 数据源 MAE RMSE MAPE/% 17:30 18:00 18:30 19:00 17:30 18:00 18:30 19:00 17:30 18:00 18:30 19:00 GCN 检测器 8.01 8.24 8.01 9.70 9.38 9.76 9.23 10.95 34.33 41.10 37.91 46.26 GCN-Seq2Seq 5.30 5.47 4.96 5.67 6.53 6.76 6.20 6.82 25.93 31.33 27.80 34.72 GCN-LSTM 在线地图 4.17 4.44 3.99 4.28 5.33 5.61 5.13 5.27 22.48 27.69 24.28 28.44 -
[1] 张建旭, 郭力玮. 基于在线地图交通态势分析的路网拥堵状态识别[J]. 交通运输系统工程与信息, 2018, 18(5): 79-85. https://www.cnki.com.cn/Article/CJFDTOTAL-YSXT201805012.htmZHANG Jian-xu, GUO Li-wei. Congestion status recognition of road network based on traffic situation analysis of online map[J]. Journal of Transportation Systems Engineering and Information Technology, 2018, 18(5): 79-85. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-YSXT201805012.htm [2] 郭力玮. 基于在线地图数据挖掘的交通运行态势预测方法研究[D]. 重庆: 重庆交通大学, 2019.GUO Li-wei. Research on traffic operation situation prediction method based on online map data mining[D]. Chongqing: Chongqing Jiaotong University, 2019. (in Chinese) [3] 陈华伟, 邵毅明, 敖谷昌, 等. 基于在线地图交通状态的关键道路动态识别方法[J]. 交通运输系统工程与信息, 2019, 19(5): 50-58. https://www.cnki.com.cn/Article/CJFDTOTAL-YSXT201905008.htmCHEN Hua-wei, SHAO Yi-ming, AO Gu-chang, et al. Dynamic identification method of critical roads based on traffic state of online map[J]. Journal of Transportation Systems Engineering and Information Technology, 2019, 19(5): 50-58. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-YSXT201905008.htm [4] LECUN Y, BENGIO Y, HINTON G. Deep learning[J]. Nature, 2015, 521: 436-444. doi: 10.1038/nature14539 [5] MA Xiao-lei, TAO Zhi-min, WANG Yin-hai, et al. Long short-term memory neural network for traffic speed prediction using remote microwave sensor data[J]. Transportation Research Part C: Emerging Technologies, 2015, 54: 187-197. doi: 10.1016/j.trc.2015.03.014 [6] 刘昶. 基于时间序列的经典模型和LSTM模型的城市宏观行程速度预测研究[D]. 北京: 北京交通大学, 2019.LIU Chang. Research on urban macro travel speed prediction based on classical models and LSTM model in time series models[D]. Beijing: Beijing Jiaotong University, 2019. (in Chinese) [7] 陈凯勋. 基于注意力机制的短时交通流速度预测模型研究[D]. 广州: 华南理工大学, 2019.CHEN Kai-xun. Study on short-term traffic flow velocity prediction model based on attention mechanism[D]. Guangzhou: South China University of Technology, 2019. (in Chinese) [8] ZHAO Jian-dong, GAO Yuan, YANG Zhen-zhen, et al. Truck traffic speed prediction under non-recurrent congestion: based on optimized deep learning algorithms and GPS data[J]. IEEE Access, 2019, 7: 9116-9127. doi: 10.1109/ACCESS.2018.2890414 [9] JEONG M, LEE T, JEON S, et al. Highway speed prediction using gated recurrent unit neural networks[J]. Applied Sciences, 2021, 11(7): 3059-3073. doi: 10.3390/app11073059 [10] MA Xiao-lei, DAI Zhuang, HE Zheng-bing, et al. Learning traffic as images: a deep convolutional neural network for large-scale transportation network speed prediction[J]. Sensors, 2017, 17(4): 818-833. doi: 10.3390/s17040818 [11] LIU Qing-chao, WANG Bo-chen, ZHU Yu-quan. Short-term traffic speed forecasting based on attention convolutional neural network for arterials[J]. Computer-Aided Civil and Infrastructure Engineering, 2018, 33(11): 999-1016. doi: 10.1111/mice.12417 [12] WANG Jing-yuan, GU Qian, WU Jun-jie, et al. Traffic speed prediction and congestion source exploration: a deep learning method[C]//IEEE. International Conference on Data Mining. New York: IEEE, 2016: 499-508. [13] JO D, YU B, JEON H, et al. Image-to-image learning to predict traffic speeds by considering area-wide spatio-temporal dependencies[J]. IEEE Transactions on Vehicular Technology, 2019, 68(2): 1188-1197. doi: 10.1109/TVT.2018.2885366 [14] 唐克双, 陈思曲, 曹喻旻, 等. 基于Inception卷积神经网络的城市快速路行程速度短时预测[J]. 同济大学学报: 自然科学版, 2021, 49(3): 370-381. https://www.cnki.com.cn/Article/CJFDTOTAL-TJDZ202103009.htmTANG Ke-shuang, CHEN Si-qu, CAO Yu-min, et al. Short-term travel speed prediction for urban expressways based on convolutional neural network with inception module[J]. Journal of Tongji University: Natural Science, 2021, 49(3): 370-381. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-TJDZ202103009.htm [15] YU Hai-yang, WU Zhi-hai, WANG Shu-qin, et al. Spatio-temporal recurrent convolutional networks for traffic prediction in transportation networks[J]. Sensors, 2017, 17(7): 1501-1516. doi: 10.3390/s17071501 [16] ZANG Di, LING Jia-wei, WEI Zhi-hua, et al. Long-term traffic speed prediction based on multiscale spatio-temporal feature learning network[J]. IEEE Transactions on Intelligent Transportation Systems, 2019, 20(10): 3700-3709. doi: 10.1109/TITS.2018.2878068 [17] 杨建喜, 郁超顺, 李韧, 等. 基于多周期组件时空神经网络的路网通行速度预测[J]. 交通运输系统工程与信息, 2021, 21(3): 112-119. https://www.cnki.com.cn/Article/CJFDTOTAL-YSXT202103014.htmYANG Jian-xi, YU Chao-shun, LI Ren, et al. Traffic network speed prediction via multi-periodic-component spatial-temporal neural network[J]. Journal of Transportation Systems Engineering and Information Technology, 2021, 21(3): 112-119. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-YSXT202103014.htm [18] WANG Jia-wei, CHEN Rui-xiang, HE Zhao-cheng. Traffic speed prediction for urban transportation network: a path based deep learning approach[J]. Transportation Research Part C: Emerging Technologies, 2019, 100: 372-385. doi: 10.1016/j.trc.2019.02.002 [19] GU Yuan-li, LU Wen-qi, QIN Ling-qiao, et al. Short-term prediction of lane-level traffic speeds: a fusion deep learning model[J]. Transportation Research Part C: Emerging Technologies, 2019, 106: 1-16. doi: 10.1016/j.trc.2019.07.003 [20] LYU Zhong-jian, XU Jia-jie, ZHENG Kai, et al. LC-RNN: a deep learning model for traffic speed prediction[C]//IJCAI. International Joint Conference on Artificial Intelligence. Stockholm: IJCAI, 2018: 3470-3476. [21] YU B, LEE Y, SOHN K. Forecasting road traffic speeds by considering area-wide spatio-temporal dependencies based on a graph convolutional neural network (GCN)[J]. Transportation Research Part C: Emerging Technologies, 2020, 114: 189-204. doi: 10.1016/j.trc.2020.02.013 [22] GE Liang, LI Hang, LIU Jun-ling, et al. Temporal graph convolutional networks for traffic speed prediction considering external factors[C]//IEEE. International Conference on Mobile Data Management. New York: IEEE, 2019: 234-242. [23] ZHANG Zheng-chao, LI Meng, LIN Xi, et al. Multistep speed prediction on traffic networks: a deep learning approach considering spatio-temporal dependencies[J]. Transportation Research Part C: Emerging Technologies, 2019, 105: 297-322. doi: 10.1016/j.trc.2019.05.039 [24] XIE Zhi-pu, LYU Wei-feng, HUANG Shang-fo, et al. Sequential graph neural network for urban road traffic speed prediction[J]. IEEE Access, 2019, 8: 63349-63358. http://ieeexplore.ieee.org/document/8708297 [25] LU Zhi-long, LYU Wei-feng, CAO Ya-bin, et al. LSTM variants meet graph neural networks for road speed prediction[J]. Neurocomputing, 2020, 400: 34-45. doi: 10.1016/j.neucom.2020.03.031 [26] DEFFERRARD M, BRESSON X, VANDERGHEYNST P. Convolutional neural networks on graphs with fast localized spectral filtering[C]//NIPS. International Conference on Neural Information Processing Systems. Barcelona: NIPS, 2016: 3844-3852. [27] HOCHREITER S, SCHMIDHUBER J. Long short-term memory[J]. Neural Computation, 1997, 9(8): 1735-1780. doi: 10.1162/neco.1997.9.8.1735 [28] 苏飞, 董宏辉, 贾利民, 等. 基于时空相关性的城市交通路网关键路段识别[J]. 交通运输系统工程与信息, 2017, 17(3): 213-221. https://www.cnki.com.cn/Article/CJFDTOTAL-YSXT201703032.htmSU Fei, DONG Hong-hui, JIA Li-min, et al. Identification of critical section in urban traffic road network based on space-time correlation[J]. Journal of Transportation Systems Engineering and Information Technology, 2017, 17(3): 213-221. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-YSXT201703032.htm [29] LECUN Y, BOSER B, DENKER J S, et al. Backpropagation applied to handwritten zip code[J]. Neural Computation, 1989, 1(4): 541-551. doi: 10.1162/neco.1989.1.4.541 [30] HOPFIELD J J. Neural networks and physical systems with emergent collective computational abilities[J]. Proceedings of the National Academy of Sciences of the United States of America, 1982, 79(8): 2554-2561. doi: 10.1073/pnas.79.8.2554