原文传递 STATISTICAL DATA FILTERING AND AGGREGATION TO HOUR TOTALS OF INTELLIGENT TRANSPORTATION SYSTEM 30-S AND 5-MIN VEHICLE COUNTS.
题名: STATISTICAL DATA FILTERING AND AGGREGATION TO HOUR TOTALS OF INTELLIGENT TRANSPORTATION SYSTEM 30-S AND 5-MIN VEHICLE COUNTS.
作者: Schmoyer-R; Hu-PS; Goeltz-RT
关键词: Data-aggregation; Data-archiving; Data-filtering; Feasibility-analysis; Intelligent-transportation-systems; Lane-occupancy; Long-Island-New-York; Orlando-Florida; Traffic-counts; Traffic-data; Traffic-speed
摘要: For traffic analysts who rely on traditional traffic monitoring data from state and local transportation departments, Intelligent Transportation System (ITS) count, speed, and occupancy data are a largely untapped resource. As with most data, however, ITS data are not immune to outliers, missing values, and other anomalies, which can be time-consuming and costly in data analysis. Furthermore, because ITS data are recorded at intervals as short as 20 s, the tasks of archiving and disseminating them are substantial. Therefore, before ITS data are applied in traditional traffic analysis problems, and before archiving and disseminating them, their fitness for traditional uses should be demonstrated, and methods should be established for ensuring their fitness. Traffic counts from ITS installations in Orlando, Florida, and Long Island, New York were examined. Methods were developed for (a) filtering out data values beyond credible limits and runs of the same value too long to be credible and (b) dealing with zeros, which, ambiguously, indicate either missing values or actual zero counts. Because hour totals are often the focus in traditional traffic analyses, methods were also developed for aggregating the remaining counts to estimates of hour totals. Adjustments were made for the joint effect of missing values and trends within hours. Application of the methods to the Florida and Long Island ITS traffic counts and comparisons with hour counts from state departments of transportation demonstrated the feasibility of using ITS counts in traditional traffic analyses.
总页数: Transportation Research Record. 2001. (1769) pp79-86 (6 Fig., 3 Tab., 2 Ref.)
报告类型: 科技报告
相关文献
检索历史
应用推荐