TRAVEL TIME DATA COLLECTION FOR MEASUREMENT OF ADVANCED TRAVELER INFORMATION SYSTEMS ACCURACY
TRAVEL TIME DATA COLLECTION FOR MEASUREMENT OF ADVANCED TRAVELER INFORMATION SYSTEMS ACCURACY
复制标题
用于测量高级旅行者信息系统准确性的旅行时间数据收集
DOI：
--
复制
发表时间：
2004
期刊：
影响因子：
--
通讯作者：
D. Wunderlich
中科院分区：
文献类型：
--
作者：
A. Toppen;D. Wunderlich
实时交通信息的用户想要知道他们的行程需要多长时间，以便在替代路线或模式之间进行选择，确定何时离开，或者在必要时调整他们的时间表。这引起了交通管理者的兴趣，他们将点到点的旅行时间作为高级旅行者信息系统（ATIS）的一部分来估计。当然，不可能总是以完美的精度预测点对点的旅行时间。存在许多潜在误差的来源，包括传感器的可靠性、根据传感器测量值计算行程时间，以及无法准确预测在即将进行的行程过程中条件将如何变化。在本文中，我们强调测量ATIS旅行时间估计的准确性的重要性，讨论不同的数据收集技术的利弊，并提供足够的研究成本估计。这可能就像驾驶一辆装备适度的探测车来测量“地面实况”旅行时间一样简单。探测车技术是评估ATIS准确性的最佳方法，ATIS覆盖了城域网中的多个网段。我们估计，100个探测车运行将包括一个平均规模的大都市区的足够的研究。收集这么多数据将花费大约21 000美元。日常旅行时间的变化是一个关键指标，如何准确的ATIS旅行时间估计需要。20%的误差是一个合适的初始目标，尽管这个值可能会因大都市地区而显着变化。在理想情况下，可以使用存档的ATIS旅行时间估计来计算网络范围的可变性。然而，如果这些估计被证明是不准确的基础上获得的地面实况数据从探测器的研究，这将导致一个不准确的估计变异性。因此，如果行程时间估计误差为20%或更差，则应采用使用车牌匹配技术的额外现场数据，以准确表征每日变化。就一项研究而言，估计费用约为48 000美元。
Users of real-time traffic information want to know how long their trips are going to take in order to choose between alternate routes or modes, determine when to leave, or adjust their schedules if necessary. This has spurred interest among traffic managers to estimate point-to-point travel times as part of Advanced Traveler Information Systems (ATIS). Of course, it is impossible to always predict point-to-point travel times with perfect accuracy. There are numerous sources of potential error including the reliability of sensors, the calculation of travel time from sensor measurements, and the inability to accurately forecast how conditions will change over the course of a pending trip. In this paper we underscore the importance of measuring the accuracy of ATIS travel time estimates, discuss the pros and cons of different data collection techniques and provide cost estimates for sufficient studies. This may be as simple as driving a moderately-equipped probe vehicle to measure "ground truth" travel times. Probe vehicle techniques are the best approach to assess the accuracy of an ATIS that covers a number of segments in a metropolitan network. We estimate that 100 probe vehicle runs would comprise a sufficient study for an average sized metropolitan area. Collecting this much data would cost approximately $21,000. Day-to-day travel time variability is a key indicator for how accurate ATIS travel time estimates need to be. An error of 20% is a suitable initial target, though this value may vary significantly by metropolitan area. Under ideal circumstances, one could calculate network-wide variability using archived ATIS travel time estimates. However, if these estimates are shown to be inaccurate based on the ground truth data obtained from the probe vehicle study, this would lead to an inaccurate estimate of variability. Therefore, if travel time estimation error is 20% or worse, additional field data using license plate matching techniques should be taken for the purpose of accurately characterizing day-today variability. For a single study, it is estimated that this would cost approximately $48,000.
DOI：
--
发表时间：
2000-09
期刊：
--
影响因子：
--
作者：
J. Ygnace;C. Drane;Y. Yim;Renaud de Lacvivier
通讯作者：
J. Ygnace;C. Drane;Y. Yim;Renaud de Lacvivier