An Expressway Traffic Flow Prediction Method Considering Inter-Lane Differences and Upstream and Downstream Cross-Section Correlations
-
摘要: 在现有的交通流量预测研究中，并未充分考虑断面道路内不同车道间的交通流量差异性以及上下游断面交通流量相关性。研究了结合主成分分析法(principal component analysis，PCA)与长短期记忆神经网络(long short-term memory，LSTM)的快速路交通流量预测框架，可以满足智能网联技术实时性和准确性的需求。收集城市快速路的交通流量数据，应用快速傅里叶变换方法(fast fourier transform，FFT)进行数据预处理，以提高原始数据的可预测性能；通过PCA方法对车道间的横向及纵向交通流量进行特征融合，建立车道间交通流量的关联性数据，以降低数据维度；并将关联性数据融入到LSTM模型中，进行车道级交通流量预测并汇总其预测结果，得到断面交通流量的预测值。选取武汉市三环线上的城市快速路卡口检测数据对本文方法进行验证。结果表明：考虑车道间差异和上下游断面关联的模型能够提高断面交通流量的预测精度，相较于仅考虑时间特征的断面交通流量预测结果，平均绝对误差、均方根误差和平均绝对百分比误差分别能降低6.66%，6.23%，17.51%；与单独考虑上下游断面关联性或者车道间差异的断面交通流量预测结果相比均具有更好的预测效果，在平均绝对误差、均方根误差和平均绝对百分比误差上的优化幅度，最低可降低1.53%，最高可降低12.88%；此外，所提的模型相较于支持向量机回归(support vector regression，SVR)和随机森林(random forest，RF)算法具有更高的预测精度；并且在分时段预测中，在晚高峰和平峰时段预测精度表现更佳。Abstract: In the existing traffic flow prediction studies, the differences of traffic flow between different lanes within a cross-section road and the correlation of traffic flow between upstream and downstream cross-sections are not fully considered. This research develops a framework for expressway traffic flow prediction by integrating principal component analysis (PCA) and long short-term memory (LSTM) neural networks, which satisfies the real-time and accuracy requirements of intelligent connected technologies. The traffic flow data of urban expressways are collected, and the fast Fourier transform (FFT) method is applied for data preprocessing to improve the predictability of the original data. PCA is used to fuse the features of lateral and longitudinal traffic flow between lanes, establishing correlated traffic flow data between lanes to reduce data dimensionality. The correlated data is then integrated into the LSTM model for lane-level traffic flow prediction, and the prediction results aggregated to obtain the cross-section traffic flow estimates. The proposed method is validated using traffic data from expressway checkpoint detectors along the Third Ring Road in Wuhan. The results demonstrate that the model considering inter-lane differences and upstream and downstream cross-section correlations can improve the accuracy of cross-section traffic flow prediction. Compared to prediction results that only consider temporal features, the mean absolute error, root mean square error, and mean absolute percentage error can be reduced by 6.66%, 6.23%, and 17.51%, respectively. When compared with models that account for either upstream-downstream correlations or inter-lane differences alone, the proposed model demonstrates superior performance, with reductions in mean absolute error, root mean square error, and mean absolute percentage error ranging from a minimum of 1.53% to a maximum of 12.88%. Additionally, the proposed model demonstrates higher prediction accuracy than support vector regression (SVR) and random forest (RF) algorithms. In time-segmented predictions, the model performs particularly well during the evening peak and off-peak hours.
-
图 4 在不同时间段上不考虑交通流量关联性的断面交通流量预测结果
Figure 4. Prediction results of cross-sectional traffic flow without considering traffic flow correlation at different time periods
图 5 在不同时间段上考虑交通流量关联性的断面交通流量预测结果
Figure 5. Prediction results of cross-sectional traffic flow considering traffic flow correlation at different time periods
表 1 米粮路断面交通流量的PCA提取
Table 1. PCA extraction of longitudinal traffic flow between lanes
特征提取 KMO 累计贡献率/% 与上下游断面交通流量提取 0.763*** 96.814 断面内横向车道间交通流量提取 0.779*** 95.237 车道1纵向交通流量提取 0.753*** 95.995 车道2纵向交通流量提取 0.758*** 96.027 车道3纵向交通流量提取 0.752*** 95.406 注：“***”为经过bartlett检验，P < 0.001，说明数据是适合主成分分析的。 
下载: 导出CSV
表 2 不考虑交通流量关联性的断面交通流量预测结果
Table 2. Prediction results of cross-sectional traffic flow without considering traffic flow correlation
实验方案 MAE RMSE MAPE/% SVR-未考虑车道间差异 20.698 7 28.010 5 27.981 1 RF-未考虑车道间差异 16.763 2 22.190 9 30.020 9 LSTM-未考虑车道间差异 12.296 7 15.801 2 18.566 3 LSTM-考虑车道间差异 11.746 0 15.278 5 17.578 8
下载: 导出CSV
表 3 考虑交通流量关联性的断面交通流量预测结果
Table 3. Prediction results of cross-sectional traffic flow considering traffic flow correlation
实验方案 MAE RMSE MAPE/% SVR-未考虑车道间差异 18.164 1 23.437 0 25.986 0 RF-未考虑车道间差异 14.723 2 19.495 3 23.932 3 LSTM-未考虑车道间差异 11.715 9 15.046 7 17.491 8 LSTM-考虑车道间差异 11.477 9 14.816 9 15.315 3
下载: 导出CSV
表 4 在不同时间段上不考虑交通流量关联性的断面交通流量预测误差
Table 4. Error in predicting cross-sectional traffic flow without considering traffic flow correlation at different time periods
误差 早高峰 晚高峰 平峰 MAE RMSE MAPE/% MAE RMSE MAPE/% MAE RMSE MAPE/% SVR-未考虑车道间差异 47.558 6 55.641 1 40.708 2 27.478 9 34.470 9 23.540 4 15.236 8 19.234 8 12.140 6 RF-未考虑车道间差异 14.001 3 16.643 4 8.561 1 22.487 0 28.051 3 16.499 2 15.970 1 18.413 5 12.515 1 LSTM-未考虑车道间差异 21.470 0 25.583 4 15.157 5 13.213 9 17.115 4 8.420 5 11.839 8 15.407 7 8.215 6 LSTM-考虑车道间差异 23.299 7 27.904 5 17.310 3 12.973 7 16.332 0 8.614 9 10.926 8 13.678 9 7.849 1
下载: 导出CSV
表 5 在不同时间段上考虑交通流量关联性的断面交通流量预测误差
Table 5. Error in predicting cross-sectional traffic flow considering traffic flow correlation at different time periods
误差 早高峰 晚高峰 平峰 MAE RMSE MAPE/% MAE RMSE MAPE/% MAE RMSE MAPE/% SVR-未考虑车道间差异 22.395 1 25.877 2 14.887 3 24.344 1 31.625 7 20.144 4 17.728 2 20.633 7 14.618 2 RF-未考虑车道间差异 17.507 0 22.106 8 11.015 8 16.555 9 21.358 8 12.021 9 10.473 5 14.257 0 7.768 2 LSTM-未考虑车道间差异 19.436 3 23.761 3 13.273 5 14.842 5 17.884 8 9.363 4 11.807 1 15.751 5 8.177 6 LSTM-考虑车道间差异 22.464 4 26.580 0 16.544 6 14.755 6 17.844 5 10.048 3 10.788 1 12.613 2 8.062 2
下载: 导出CSV
-
[1] YANG B, SUN S, LI J, et al. Traffic flow prediction using LSTM with feature enhancement[J]. Neurocomputing, 2019, 332: 320-327. doi: 10.1016/j.neucom.2018.12.016 [2] 王殿海, 蔡正义, 曾佳棋, 等. 城市交通控制中的数据采集研究综述[J]. 交通运输系统工程与信息, 2020, 20(3): 95-102.WANG D H, CAI Z Y, ZENG J Q, et al. Review of traffic data collection research on urban traffic control[J]. Journal of Transportation Systems Engineering and Information Technology, 2020, 20(3): 95-102. (in Chinese) [3] LI M, LI M, LIU B, et al. Spatio-temporal traffic flow prediction based on coordinated attention[J]. Sustainability, 2022, 14 (12): 7394. doi: 10.3390/su14127394 [4] 陈如清, 李嘉春, 俞金寿. 基于FWADE-ELM的短时交通流预测方法[J]. 控制与决策, 2021, 36(4): 925-932.CHEN R Q, LI J C, YU J S. Short-term traffic flow forecasting based on hybrid FWADE-ELM[J]. Control and Decision, 2021, 36(4): 925-932. (in Chinese) [5] 陆文琦, 芮一康, 冉斌, 等. 智能网联环境下基于混合深度学习的交通流预测模型[J]. 交通运输系统工程与信息, 2020, 20(3): 47-53.LU W Q, RUI Y K, RAN B, et al. Traffic flow prediction based on hybrid deep learning under connected and automated vehicle environment[J]. Journal of Transportation Systems Engineering and Information Technology, 2020, 20(3): 47-53. (in Chinese) [6] 赵恒辉, 黄德启, 曾蓉, 等. 基于时空注意力Bi-LSTM模型的短时交通流预测[J]. 计算机仿真, 2022, 39(9): 177-181.ZHAO H H, HUANG D Q, ZENG R, et al. Short term traffic flow prediction based on spatial-temporal attention Bi-LSTM model[J]. Computer Simulation, 2022, 39(9): 177-181. (in Chinese) [7] LU S, ZHANG Q, CHEN G, et al. A combined method for short-term traffic flow prediction based on recurrent neural network[J]. Alexandria Engineering Journal, 2021, 60(1): 87-94. doi: 10.1016/j.aej.2020.06.008 [8] 陈宇, 王炜, 华雪东, 等. 基于递归框架的高速公路交通流量实时预测方法[J]. 交通信息与安全, 2023, 41(1): 124-131. doi: 10.3963/j.jssn.1674-4861.2023.01.013CHEN Y, WANG W, HUA X D, et al. A recursive framework-based approach for real-time traffic flow[J]. Journal of Transport Information and Safety, 2023, 41(1): 124-131. (in Chinese) doi: 10.3963/j.jssn.1674-4861.2023.01.013 [9] MA D, SONG X, LI P. Daily traffic flow forecasting through a contextual convolutional recurrent neural network modeling inter-and intra-day traffic patterns[J]. IEEE Transactions on Intelligent Transportation Systems, 2020, 22(5): 2627-2636. [10] 温惠英, 张东冉, 陆思园. GA-LSTM模型在高速公路交通流预测中的应用[J]. 哈尔滨工业大学学报, 2019, 51(9): 81-87.WEN H Y, ZHANG D R, LU S Y. Application of GA-LSTM model in highway traffic flow prediction[J]. Journal of Harbin Institute of Technology, 2019, 51(9): 81-87. (in Chinese) [11] 王殿海, 谢瑞, 蔡正义. 基于最优汇集时间间隔的城市间断交通流预测[J]. 浙江大学学报(工学版), 2023, 57(8): 1607-1617.WANG D H, XIE R, CAI Z Y. Prediction of urban interrupted traffic flow based on optimal convergence time interval[J]. Journal of Zhejiang University(Engineering Science), 2023, 57(8): 1607-1617. (in Chinese) [12] ZHENG Y, LI W, ZHENG W, et al. Lane-level heterogeneous traffic flow prediction: a spatiotemporal attention-based encoder-decoder model[J]. IEEE Intelligent Transportation Systems Magazine, 2023, 15(3): 51-67. [13] ZHOU J, SHUAI S, WANG L, et al. Lane-level traffic flow prediction with heterogeneous data and dynamic graphs[J]. Applied Sciences, 2022, 12(11): 5340. [14] 侯越, 崔菡珂, 邓志远. 横向相关性及参数影响下的车道级交通预测[J]. 公路交通科技, 2022, 39(5): 122-130.HOU Y, CUI H K, DENG Z Y. Lane level traffic prediction under influence of lateral correlation and parameters[J]. Journal of Highway and Transportation Research and Development, 2022, 39(5): 122-130. (in Chinese) [15] 侯越, 郑鑫, 韩成艳. 一种融合纵横时空特征的交通流预测方法[J]. 西安电子科技大学学报, 2023, 50(5): 65-74.HOU Y, ZHENG X, HAN C Y. Traffic flow prediction method for integrating longitudinal and horizontal spatiotemporal characteristics[J]. Journal of Xidian University, 2023, 50 (5): 65-74. (in Chinese) [16] 周军勇, 石雪飞, 阮欣, 等. 高速公路分车道荷载差异及其响应特性[J]. 同济大学学报(自然科学版), 2018, 46(4): 458-464.ZHOU J Y, SHI X F, RUAN X, et al. Land load disparities and their loading effect characteristics of freeway[J]. Journal of Tongji University(Natural Science), 2018, 46(4): 458-464. (in Chinese) [17] 阮欣, 周可攀, 周军勇. 某八车道高速公路车流特性及荷载效应[J]. 同济大学学报(自然科学版), 2015, 43(4): 555-561.RUAN X, ZHOU K P, ZHOU J Y. Vehicle flow characteristics and load effect of a eight-lane highway[J]. Journal of Tongji University(Natural Science), 2015, 43(4): 555-561. (in Chinese) [18] 张惠玲, 奚邦顺. 老年人过街比例与信号交叉口行人过街速度设置研究[J]. 交通运输系统工程与信息, 2021, 21 (1): 214-220.ZHANG H L, XI B S. Signalized intersection pedestrian crossing design speed and ederly pedestrian proportion relationship study[J]. Journal of Transportation Systems Engineering and Information Technology, 2021, 21(1): 214-220. (in Chinese) [19] SINGH N K, TANGIRALAA K, VANAJAKSHI L D. A multivariate analysis framework for vehicle detection from loop data under heterogeneous and less lane disciplined traffic[J]. IEEE Access, 2021, (9): 143580-143591. [20] 魏晓悦, 靳春玲, 贡力, 等. 基于PCA-改进RBF神经网络模型的铁路隧道突水风险评价[J]. 铁道科学与工程学报, 2021, 18(3): 794-802.WEI X Y, JIN C L, GONG L, et al. Risk evaluation of railway tunnel water inrush based on PCA-improved RBF neural network model[J]. Journal of Railway Science and Engineering, 2021, 18(3): 794-802. (in Chinese) [21] ARORA S, KESHARI A K. Pattern recognition of water quality variance in Yamuna River(India)using hierarchical agglomerative cluster and principal component analyses[J]. Environmental Monitoring and Assessment, 2021, 193: 494. [22] 李岩, 王泰州, 徐金华, 等. 面向动态交通分配的交通需求深度学习预测方法[J]. 交通运输系统工程与信息, 2024, 24(1): 115-123.LI Y, WANG T Z, XU J H, et al. Traffic demand prediction method based on deep learning for dynamic traffic assignment[J]. Journal of Transportation Systems Engineering and Information Technology, 2024, 24(1): 115-123. (in Chinese) [23] 张璐, 冯东明, 吴刚. 基于LSTM网络的车辆动态荷载识别方法[J]. 东南大学学报(自然科学版), 2023, 53(2): 187-192.ZHANG L, FENG D M, WU G. Dynamic vehicle load identification method based on LSTM network[J]. Journal of Southeast University(Natural Science Edition), 2023, 53 (2): 187-192. (in Chinese) -
点击查看大图
图(5) / 表(5)
计量
- 文章访问数: 36
- HTML全文浏览量: 14
- PDF下载量: 6
- 被引次数: 0