交通运输工程与信息学报

居民出行调查是科学制定城市交通发展战略、政策、技术法规的基础性工作,在日益复杂的城市交通环境中,提高数据扩样的精度一直是一个研究重点。本文按照数据的扩样流程系统阐述了数据扩样方法。首先,分层次的直接扩样方法有效地控制了扩样后数据的分布特征;其次,根据核查线交通量等客观数据,基于最大似然法原理,应用交通规划软件CUBE的矩阵估算模块对数据进行了核查校正;最后,针对抽样技术的缺陷,增加中区校核的过程,对校核到的沉默需求分四种方法返回出行目的进行比较寻优。扩样方法在2005年广州居民出行调查中得到了应用,取得了较好的效果。

Resident trip survey (RTS) is a fundamental work for making urban traffic development strategy, policy and technological regulation scientifically. The research emphasis is on improving the precision of data sampling expansion under more complicated urban traffic circumstances. This paper described the data sampling expansion method systematically. First of all,the direct sampling expansion methods in various hierarchies controlled the distribution characteristic of the data process effectively; Secondly , according to the traffic volume of screen line,the verified data was through the matrix estimation module of the traffic planning software CUBE based on the maximum likelihood principle. Finally,it added a verified course of the middle traffic zones for the defects of the technology; the verified underreport demand had been compared and optimized based on trip purpose with four kinds of methods. The data sampling expansion method has been used in 2005,s RTS of Guangzhou and made a good result.