常规公交车辆串车形成及预测建模

发布时间：2018-06-18 09:48

本文选题：公共交通 + 串车　；参考：《东南大学学报(自然科学版)》2017年06期

【摘要】：首先结合车头时距不稳定性的概念,对串车现象给出判定依据;分析串车问题产生的6种典型过程,建立串车问题数学模型,对串车发生的原因进行理论分析.然后,以自贡市38路公交线路实际数据为例,采用方差分析与回归分析筛选出7个串车形成影响因素,并对其重要性进行排序.最后,建立基于径向基函数神经网络的串车问题预测模型,对未来站的车头时距进行预测,并选取自贡市38路公交车的实时数据进行模型验证.在38 000余组数据中随机选择300组数据进行训练,对比30组测试数据.结果表明,学习得到的预测值与实际值偏差10%以内的样本点占90%,结果良好,证明了所建立的模型具有较好的适用性.
[Abstract]:Firstly, combining the concept of head-length instability, the paper gives the judgment basis for the phenomenon of train string, analyzes the six typical processes of the train string problem, establishes the mathematical model of the train string problem, and makes a theoretical analysis on the cause of the train string problem. Then, taking the actual data of 38 bus routes in Zigong City as an example, seven factors affecting the formation of buses are selected by means of ANOVA and regression analysis, and their importance is ranked. Finally, the prediction model of train string problem based on radial basis function neural network is established to predict the head-length of the future station, and the real-time data of No. 38 bus in Zigong city are selected to verify the model. 300 groups of data were randomly selected for training in more than 38 000 groups of data and 30 groups of test data were compared. The results show that the sample points within 10% of the deviation between the predicted values and the actual values are 90%, and the results are good, which proves that the proposed model has good applicability.
【作者单位】：东南大学城市智能交通江苏省重点实验室;东南大学现代城市交通技术江苏高校协同创新中心;东南大学江苏省物联网技术与应用协同创新中心;东南大学物联网交通应用研究中心;
【基金】：国家重点研发计划资助项目(2016YFB0100906) 国家自然科学基金资助项目(61620106002,51308115)
【分类号】：U491.17

【相似文献】