详情

原文传递 STATISTICAL MODELS OF AT-GRADE INTERSECTION ACCIDENTS - ADDENDUM.

题名：	STATISTICAL MODELS OF AT-GRADE INTERSECTION ACCIDENTS - ADDENDUM.
作者：	Bauer-KM; Harwood-DW
关键词：	At-grade-intersections; Average-daily-traffic; Four-leg-intersections; Geometric-design; Lognormal-regression-analysis; Mathematical-models; Multiple-vehicle-accidents; Negative-binomial-regression-analysis; Poisson-regression-analysis; Regression-analysis; Rural-areas; Signalized-intersections; Single-vehicle-accidents; Stop-controlled-intersections; Three-leg-intersections; Traffic-accidents; Traffic-volume; Urban-areas
摘要：	This report is an addendum to the work published in FHWA-RD-96-125 titled "Statistical Models of At-Grade Intersection Accidents." The objective of both research studies was to develop statistical models of the relationship between traffic accidents and highway geometric elements for at-grade intersections. While the previously published report used only multiple-vehicle accidents in developing predictive models, this addendum presents models based on all collision types (including both multiple-vehicle and single-vehicle accidents). The statistical modeling approaches used in the research included lognormal, Poisson, and negative binomial regression analyses. The models for all collision types are similar to those developed in the previous report for multiple-vehicle accidents. The regression models of the relationships between accidents and intersection geometric design, traffic control, and traffic volume variables were found to explain between 16 and 39% of the variability in the accident data. However, most of that variability was explained by the traffic volume variables considered (major road and crossroad average daily traffic). Geometric design variables accounted for only a small additional portion of the variability. Generally, negative binomial regression models were developed to fit the accident data at rural, three- and four-leg, STOP-controlled intersections and urban, three-leg, STOP-controlled intersections. On the other hand, lognormal regression models were found more appropriate for modeling accidents at urban, four-leg, STOP-controlled and urban, four-leg, signalized intersections. The decision to use negative binomial or lognormal regression analysis was based on evaluation of the accident frequency distribution for the specific categories of intersections.
总页数：	2000/03. 9806-9810 pp68 (10 Fig., 26 Tab., Refs., 1 App.)
报告类型：	科技报告