00815915

Component

https://doi.org/10.3141/1746-03

http://scholar.google.com...mond&publication_year=2001

Access management can improve safety or operations but may reduce accessibility for adjacent residences or businesses, causing conflict between the needs of local and through traffic and resulting in access management decisions being challenged by the public. Accordingly, transportation administrators require assurance regarding the accuracy of projected safety impacts of access management decisions. Several existing mathematical models quantify these impacts for selected access management techniques. Since new models require substantial resources to construct, it is prudent to investigate the extent to which existing models can be applied in other locations. A study of mathematical models that predict the number of crashes that will result as a function of signal spacing, median treatments, unsignalized driveways, and other key variables is reported. The study compared five models using a 10-year set of geometric, operational, and crash data for three case study corridors. The crash-prediction ability of these models was assessed using traffic flow data other than the data set used to construct the models. Because some of the models were not developed for application elsewhere or were intended to be used only with a site-specific adjustment, the error percentages indicate the extent to which the models are transferable, but they do not constitute a critique of the previous research. With no site-specific adjustment, the accuracy of the models ranges from an average error of 34% to a few hundred percent. With simple site-specific adjustments, errors are in the range of 27% to 29%. Substantially less effort is required for these site-specific adjustments than would be required to create a new mathematical model. Recommendations were developed for using these models in practice, understanding their limitations, and interpreting their sensitivity to key inputs, data needs, and computational requirements.

This paper appears in Transportation Research Record No. 1746, Highway Safety: Modeling, Analysis, Management, Statistical Methods, and Crash Location.

English

p. 14-21

2001

Transportation Research Record

0309072042

Figures (3) ; References (12) ; Tables (5)

Access control (Transportation); Accuracy; Case studies; Decision making; Highway safety; Mathematical models; Mathematical prediction; Recommendations; Traffic crashes

Site-specific adjustments; Transferability

Data and Information Technology; Design; Highways; Research; Safety and Human Factors; I82: Accidents and Transport Infrastructure

TRIS, TRB, ATRI

Aug 29 2001 12:00AM

• AUTOMATED ACCIDENT DETECTION SYSTEM
• BAYESIAN IDENTIFICATION OF HIGH-RISK INTERSECTIONS FOR OLDER DRIVERS VIA GIBBS SAMPLING
• COMPUTING AND INTERPRETING ACCIDENT RATES FOR VEHICLE TYPES OR DRIVER GROUPS
• DEVELOPMENT OF ARTIFICIAL NEURAL NETWORK MODELS TO PREDICT DRIVER INJURY SEVERITY IN TRAFFIC ACCIDENTS AT SIGNALIZED INTERSECTIONS
• EVALUATION OF INEXPENSIVE GLOBAL POSITIONING SYSTEM UNITS TO IMPROVE CRASH LOCATION DATA
• GEOGRAPHIC INFORMATION SYSTEM-BASED ACCIDENT DATA MANAGEMENT FOR MEXICAN FEDERAL ROADS
• INCORPORATING CRASH RISK IN SELECTING CONGESTION-MITIGATION STRATEGIES: HAMPTON ROADS AREA (VIRGINIA) CASE STUDY
• INJURY SEVERITY IN MULTIVEHICLE REAR-END CRASHES
• INTERSTATE HIGHWAY CRASH INJURIES DURING WINTER SNOW AND NONSNOW EVENTS
• LONGITUDINAL ANALYSIS OF FATAL RUN-OFF-ROAD CRASHES, 1975 TO 1997
• SENSOR-FRIENDLY VEHICLE AND ROADWAY COOPERATIVE SAFETY SYSTEMS: BENEFITS ESTIMATION
• SIMULATION OF ROAD CRASHES BY USE OF SYSTEMS DYNAMICS