不完整车辆侦测器数据推估旅行时间之研究  

Estimation of Travel Time Based on Defective Detector

作　　者：林大杰[1] 刘霈[1] 张立言[2] 黄智建[1] 

机构地区：[1]逢甲大学,中国台中 [2]嘉义大学,中国嘉义

出　　处：《交通与计算机》2008年第4期28-30,共3页Computer and Communications

摘　　要：文中收集2007-07-12~2007-07-17日之"国道"1号北上路段111.3 K^148.9 K之车辆侦测器数据及车牌自动辨识数据(AVI),藉由相关文献了解上下游路段之特性,利用多元线性回归方式进一步探讨上下游车辆侦测器数据推估中间车辆侦测器平均车速之关系,推估出平均车速后利用Oh模式推算推估模式之旅行时间并与AVI数据比较。结果显示,在国道1号北上路段111.3 K^148.9 K上下游路段之平均车速约略存有线性关系,同时,藉由推估之旅行时间与AVI数据验证与比较,研究推估之旅行时间与完整旅行时间之平均误差百分比(MAPE)在6.874%以下,显示有良好预测;均方根误差(RMSE)在34.044以下,代表模式的可靠度尚可。The vehicle detector data from Highway 1, northbound from 111.3 Km to 148. 9 Km, dated from July 12 to July 17, 2007 was used for analysis. This study tried to understand the relationship between data from one vehicle detector and its upstream and downstream detectors, and to estimate the data from one highway location with a defective detector based on the data from its upstream and downstream detectors. The result of study indicates that the relationship between the average speed of highway vehicle flow at one location and its adjacent upstream and downstream locations is somewhat linear, and that the result of using estimated average speed data in travel time predictions is satisfactory.

关 键 词：旅行时间 辆侦测器 元线性回归 

分 类 号：U491[交通运输工程—交通运输规划与管理]