车队离散模型的最佳统计时间研究

Research on the Best Statistical Time of Platoon Dispersion Model

为对车队离散模型进行精度验证,寻找出最适合模型验证的数据的统计方法,选择对适合城市道路交叉口之间路段的Robertson提出的车队离散模型进行研究.采用多点摄像法对北京市五条路段的交通流进行实地调查,将交通流数据按照1~36s不同的时间间隔进行统计,并在不同的统计间隔下分别验证Robertson提出的车队离散模型,得出该模型在1~36s不同的统计间隔下的误差.对误差的均值和方差以及百分比进行数学分析,从而得出该模型中调查数据的最佳统计时间为6~10s,并采用T检验法进一步验证了该结论.

In order to verify the accuracy of the platoon dispersion model and find out the statistical method of data most suitable for the model validation, the platoon dispersion model proposed by Ro- bertson is studied. This model is particularly suitable for the sections between urban road intersec- tions. The traffic flows of five roads in Beijing city are investigated by multi-spot photography. The traffic flow data are collected at different time intervals ranging from 1 s to 36 s and are used to verify the platoon dispersion model proposed by Robertson for different statistical intervals. Finally, the er-rors of the model for different statistical intervals are obtained. The mean, variance and percentage of the errors are analyzed and the best statistical time of the survey data in this model is found to be 6- 10 s. The conclusion is verified by the T test.