原文传递 EFFECT OF MEDIAN TREATMENT ON URBAN ARTERIAL SAFETY: AN ACCIDENT PREDICTION MODEL.
题名: EFFECT OF MEDIAN TREATMENT ON URBAN ARTERIAL SAFETY: AN ACCIDENT PREDICTION MODEL.
作者: Bonneson-JA; McCoy-PT
关键词: MATHEMATICAL-MODELS; PREDICTIONS-; ARTERIAL-STREETS-URBAN; TRAFFIC-ACCIDENTS; MEDIAN-TREATMENTS; RAISED-CURBS; TWO-WAY-LEFT-TURN-LANES; UNDIVIDED-CROSS-SECTIONS; CALIBRATIONS-; AVERAGE-DAILY-TRAFFIC; DRIVEWAYS-; UNSIGNALIZED-INTERSECTIONS; LAND-USE; ACCIDENT-RATES; ON-STREET-PARKING
摘要: The development of a model for predicting the safety of an urban arterial street with a specified median treatment is described. The median treatments considered are raised-curb median, two-way left-turn lane (TWLTL), and undivided cross section. The model calibration was based on maximum-likelihood techniques, an assumed negative binomial distribution of the residuals, and a nonlinear relationship between accident frequency and daily traffic demand and segment length. Several conclusions were formulated on the basis of the model developed. One conclusion is that average daily traffic demand, driveway density, unsignalized public street approach density, median type, and adjacent land use are significantly correlated with accident frequency. In general, accidents are more frequent on street segments with higher traffic demands, driveway densities, or public street densities. Accidents are also more frequent when the land use is business or office as opposed to residential or industrial. The undivided cross section was shown to have a significantly higher accident frequency than the TWLTL or raised-curb median treatment when parallel parking is allowed on the undivided street. When there is no parking allowed on either type of street, the difference between the undivided, TWLTL, and raised-curb median treatments is less distinct; however, the raised-curb median treatment tends to yield the lowest accident frequency in most situations.
总页数: Transportation Research Record. 1997. (1581) pp27-36 (3 Fig., 3 Tab., 15 Ref.)
报告类型: 科技报告
检索历史
应用推荐