摘要
针对浮动车数据稳定性问题,在浮动车及其数据特性分析的基础上,设计了考虑车速置信权重的交通状态参数算法。以车速、样本数和时间为衡量基准得到置信权重,据此计算交通状态参数。在出现异常数据时进行去噪处理;在数据量不足或数据连续性不好的情况下融合历史数据及临近时段数据,以反映真实交通状态。通过编程仿真和实地实验,对优化算法进行数值分析和测试,证明该算法可有效消除异常数据波动和数据量不足的影响,对交通状态参数估计具有较高的准确度和平稳性。
Based on the analysis of characteristics of floating cars and collected data,an algorithm considering confidence of vehicle velocity is proposed to solve the instability of floating car data（FCD）.In this algorithm,traffic status parameters are calculated according to relevant confidence weights,which are estimated based on vehicle velocity,quantity of floating cars and the sampling time.This algorithm can effectively remove outliers, and impute missing values by using historical or adjacent data observations.The optimized algorithm is numerically analyzed through simulation,and is tested by field experiments,showing that the proposed algorithm is capable of smoothing abnormal data fluctuations and effectively eliminating the problems related to insufficient data,and therefore provides highly accurate estimates of traffic parameters.
出处
《交通信息与安全》
2010年第3期1-5,共5页
Journal of Transport Information and Safety
基金
国家高技术研究发展计划(863计划)项目(批准号:2008AA11Z205)资助
