论文部分内容阅读
在智能交通的相关研究中,对城市路段的历史通行速度的分析正引起广泛的关注.然而,由于路段速度数据的采集车辆不能实现对全路网的所有路段的完全覆盖,有部分路段出现了长时间的数据缺失现象,这种在同一路段上连续缺失的数据构成了一个缺失序列.现有的对缺失的速度数据的估计算法主要集中在对缺失点数据的估计,因而不适合应用于本问题.研究对这类缺失序列的估计算法,把该问题建模为序列标注问题,并使用条件随机场求解.研究了对原始速度数据的编、解码策略和与被估计路段相关的路段的选择策略,在实际的交通数据上的实验表明了本文的方法在长序列的速度数据估计问题上的有效性.
In the research of intelligent traffic, the analysis of the historical speed of urban road segment is attracting wide attention.However, because the acquisition of road speed data can not achieve the complete coverage of all road segments of the whole road network, some sections appear The data missing for a long time constitutes a missing sequence which is missing continuously on the same road segment.The existing algorithms for estimating the missing speed data mainly focus on the estimation of missing point data and thus are not suitable for application to this Problem.This paper studies the estimation algorithm of missing sequence and models the problem as sequence label problem and solves it by using conditional random field.This paper studies the coding and decoding strategy of original velocity data and the choice of the links related to the estimated link The experiments on real traffic data show that the proposed method is effective in the estimation of long series of velocity data.