| 181 | 5 | 129 |
| 下载次数 | 被引频次 | 阅读次数 |
为提高用户公交出行积极性、方便管理部门合理调度公交班次,利用大数据分析公交浮动车辆历史GPS数据,考虑不同线路、公交站点地理位置、不同驾驶员、气象情况、时间分布等多因素的影响,建立了一种基于门控循环单元(gated recurrent unit, GRU)神经网络的公交到站时间预测模型。该模型结合5 000多万条原始数据,借助分布式Hadoop集群中的Spark弹性分布式数据集进行数据清理,并运用站点匹配算法进行源数据匹配、Lasso算法优化特征选项及去除干扰。实验仿真结果表明:改进的GRU模型R-square拟合度达到94.547%,并且算法效率较传统长短期记忆(long short-term memory,LSTM)神经网络提高了近14%,为进一步提高公交到站时间的预测精度与效率提供了参考。
Abstract：In order to increase the public transportation usage and the reasonability of the bus schedule by the management department, a novel prediction model of bus arrival time is proposed. This predicting model based on gated recurrent unit(GRU) neural network, analyzed the big data of historical GPS data about floating vehicle and considers the influence of different routes, bus station location, different drivers, weather conditions, time distribution and other factors. Furthermore, combining more than 50 million pieces of raw data, the model uses Spark elastic distributed data set in distributed Hadoop cluster to clean data and site matching algorithm to match source data, Lasso algorithm to optimize feature options and remove interference. The simulation results reveal that the R-square fitting degree of the improved GRU model is 94.547% and the prediction efficiency is nearly 14% higher than that of traditional long short-term(LSTM) model. It provides a reference for further improving the accuracy and efficiency of bus arrival time prediction.
[1] SUN D H, LUO H, FU L P, et al. Predicting bus arrival time on the basis of global positioning system data[J]. Transportation Research Record, 2007(2034):62-72.
[2] PADMANABAN R P S, DIVAKAR K, VANAJAKSHI L,et al. Development of a real-time bus arrival prediction system for Indian traffic conditions[J]. IET Intelligent Transport Systems, 2010, 4(3):189-200.
[3]孙棣华，赖云波，廖孝勇，等.公交浮动车辆到站时间实时预测模型[J].交通运输工程学报，2011, 11(2):84-89.
[4] YANG J S. A study of travel time modeling via time series analysis[C]//Proceedings of the 2005 IEEE Conference on Control Applications, August 28-31, 2005, Toronto. New York:IEEE Xplore,2005:855-860.
[5]熊文华，徐建闽，林思.基于BP网络的浮动车与线圈检测数据融合模型[J].计算机仿真，2009, 26(9):235-238.
[6]熊浩，韩印.基于遗传算法：支持向量机模型的快速公交行程时间算法研究[J].物流科技，2019, 42(1):113-117.
[7]宋爽.基于SVM和Kalman滤波的公交到站时间预测方法研究[D].大连：大连海事大学，2018.
[8]张铭坤，王昕.基于GRU-RNN模型的城市主干道交通时间预测[J].北京信息科技大学学报（自然科学版），2019,34(4):30-35.
[9]张金磊，罗玉玲，付强.基于门控循环单元神经网络的金融时间序列预测[J].广西师范大学学报（自然科学版），2019, 37(2):82-89.
[10] LI J L, GAO J, YANG Y, et al. Bus arrival time prediction based on mixed model[J]. China Communications,2017, 14(5):38-47.
[11] DHIVVABHARATHI B, KUMAR B A, VANAJAKSHI L.Real time bus arrival time prediction system under Indian traffic condition[C]//Proceedings of the 2016 IEEE International Conference on Intelligent Transportation Engineering(ICITE), August 20-22, 2016, Singapore. New York:IEEE Xplore, 2016:18-22.
[12] KYUNGHYUN C, van MERRIENBOER B, GULCEHRE C, et al. Learning phrase representations using RNN encoder-decoder for statistical machine translation[C]//Proceedings of the Conference on Empirical Methods in Natural Language Processing(EMNLP), Doha, Qatar. Stroudsburg:Association for Computational Linguistics, 2014:1724-1734.
[13]李鹏程.基于LSTM神经网络的公交到站时间预测[C]//中国自动化学会控制理论专业委员会.第37届中国控制会议论文集.中国自动化学会控制理论专业委员会：中国自动化学会控制理论专业委员会，2018, 5:44-49.
[14]柳小桐. BP神经网络输入层数据归一化研究[J].机械工程与自动化，2010(3):122-123.
[15]刘玉坤，邢立国.基于TensorFlow的线性回归模拟及Python实现[J].现代计算机（专业版），2018(29):92-94.
[16]朱献超.基于梯度下降和自适应学习的高维生物数据降维可视化方法研究[D].武汉：华中师范大学，2018.
[17] WANG H, CHEN X L, LI G. Survival forests with Rsquared splitting rules[J]. Journal of Computational Biology:A Journal of Computational Molecular Cell Biology, 2018,25(4):388-395.
基本信息:
DOI：10.12194/j.ntu.20190328001
中图分类号:U491.17
引用信息:
[1]陆俊天,孙玲,施佺.基于门控循环单元神经网络的公交到站时间预测[J],2020(02):43-49.DOI:10.12194/j.ntu.20190328001.
基金信息:
国家自然科学基金项目(61771265);; 江苏省“333工程”项目(BRA2017475);; 江苏省“青蓝工程”项目;; 南通市科技计划项目(CP12017001,GY12017006)