Ongoing Research

HSIS project staff conduct an ongoing program of research. Research topics are chosen by the FHWA Office of Safety Research and Development and FHWA Office of Safety as critical to their missions. An updated list of current projects is being compiled and will be available soon.

Development of a Tool for Assessing Statistical Modeling Approaches for Crash Data Analysis

There is currently no accepted method to compare the suitability, efficiency, or effectiveness of different statistical approaches for highway crash data analysis, particularly those involving cross-sectional modeling. Only limited comparisons have been conducted to demonstrate which statistical approaches are the best choices for particular analysis types and, typically, individual researchers must exercise their own judgment when choosing among approaches.

This study will develop a tool which will allow one to assess the performance of multivariate modeling methods in capturing casual relationships between individual roadway attributes and safety. In order to do this, the research team will develop an "Artificial Realistic Dataset" (ARD)–a roadway-segment-based dataset that represents realistic roadway characteristics and realistic crash counts. The crash counts will be based on a set of (hidden) "causal rules" which reflect predetermined relationships between crash frequency/severity and specific highway geometric design features or traffic control features chosen by the modeler for testing purposes.

While an ARD could be generated for a number of different roadway types, the initial effort will concentrate on rural two-lane roadway segments because of their high crash frequencies.

Documenting the Safety Effects of Shoulder Cross Sections and Horizontal Curves

This research will document the safety effects of open vs. closed cross sections (i.e. shoulders vs. curbs and gutters), shoulders width in open cross sections, and horizontal curvature on urban and suburban arterials. In addition, researchers will develop appropriate safety-based geometric design guidance for these design elements. This project is being conducted as a companion study to NCHRP Project 3-112, which will focus on the effects of lane widths on the same class of roadways.

Key questions that will be addressed by the research include:

• How does safety performance differ between urban and suburban arterials with open cross sections with shoulders and urban and suburban arterials with closed cross sections with curbs and gutters? (Also, in speed ranges where both types of cross sections are applicable, is there a safety rational for choosing one over the other? Under what circumstances is it appropriate to change from one cross section to the other in reconstruction projects?)
• For urban and suburban arterials with open cross sections, what is the effect of shoulder width on crash frequency and severity, how does the appropriate shoulder width vary with traffic volume and speed, and how should an appropriate shoulder width be selected for specific roadways?
• How does the safety performance of horizontal curves and tangents differ on urban and suburban arterials for specific traffic volumes and speeds? How does horizontal curve radius affect crash frequency and severity? Since many horizontal curves on lower speed arterials are not superelevated, is there any indication of lane-departure or roadway-departure crash patterns related to driver loss of control that might indicate a need for modified guidance?

This research is needed because there is currently an overall lack of Highway Safety Manual and Green Book guidance to assist designers in making key decisions in the design of urban and suburban arterials. In addition, the NCHRP Report 756, Highway Safety Research Agenda: Infrastructure and Operations, identified the investigation of AASHTO Design Criteria and Design Model Research Including the Science of Safety as a top-ranked research priority. A key aspect identified for consideration as part of that priority research need was Cross Section Design Criteria for the Urban Environment.