InTrans / Feb 02, 2023

New scientific methods offer insight into roadway safety performance

A rural intersection in Iowa

It takes a significant effort to develop a new scientific method to predict the safety performance of roadways. The task of implementing the novel approach to understanding intersection crashes is no less vast.

The research developed at the Institute for Transportation (InTrans) under its Center for Transportation Research and Education (CTRE) led by researchers Zachary Hans and Hossein Naraghi is now being implemented at the Iowa Department of Transportation (DOT) and led by Naraghi.

“It takes a lot of effort to change the thought process of safety practitioners to convince them of the robustness of the new method and be able to answer tons of questions they have about the new approach,” said Naraghi, who had the unique experience of being on both sides of research development and implementation.

He added, “All in all, I think my experience of being involved in development side tremendously helped me to break the huge roadblocks on the implementation side.”

What is the research?

The CTRE project first developed an intersection database—identifying all intersections in the state—and then the intersections were assigned to categories based on similar traffic control, cross section, speed limits, and traffic volumes. From there, project members determined the number of crashes at the intersections using five years’ worth of data.

Intersection categories included in the project

Finally, they developed and refined the safety performance functions (SPFs), or models, to account for the randomness of crashes and to determine the difference between the actual crash experience and what number of crashes would be expected at similarly defined intersections and roadway segments.

An SPF is an equation used to predict the average number of crashes per year at a location as a function of exposure and, in some cases, roadway or intersection characteristics. The SPFs were developed for 11 categories of paved intersections (defined as having at least one paved approach) and for 8 categories of primary road segments. The models were also developed for all crashes and severe crash categories.

“There’s a lot we can learn from having a better understanding of how a location is actually performing compared to how we’d expect it to perform, based on these models,” said Hans, who was the principal investigator on the SPFs research project and is a research engineer with CTRE. “Additionally, the network may be screened for possible opportunities for improvement.”

Prior to the development of the SPFs and the tool, the Iowa DOT and others looking into the safety performance of intersections and segments compared the crash rate at a site with the statewide average that came from a specific sample that was not a thorough representation of all intersections or segments. In essence, says Naraghi, the SPF models contain all sites with similar characteristics and group apples in the same category to allow for an apples-to-apples comparison for sites with similar characteristics (e.g., the safety performance of a 4-leg low-speed signalized intersection is not compared with a 4-leg low-speed unsignalized intersection).

The sites where the observed number of crashes, corrected for randomness, is higher than predicted crashes offer a potential for crash reduction and thus may warrant additional study to determine whether there are countermeasures that can be implemented to improve the safety at those locations.

“The SPF models are a method that predicts the number of the crashes on a site that can be compared to the observed crashes to see whether the crashes on the desired site are overrepresented in comparison to all other sites with similar characteristics,” summarized Naraghi, who now works as a safety data analyst at the Iowa DOT.

How is the research being implemented?

Implementation of the SPF project includes identifying segments’ potential for crash reduction

Much of the work on developing the models took place in 2021. However, the work in 2022 continued to develop interactive tools to better study the state’s intersections and the opportunities to improve safety at them. Naraghi has plenty of insights into how the work is being utilized by the Iowa DOT as the project has left the research arena and moved into the implementation phase.

Naraghi said each of the crash sites are ranked with high, medium, and negligible potential for crash reduction, and the ranking tiers are utilized by the Iowa DOT and other transportation agencies to identify and prioritize their safety projects.

“The SPFs are an integral part of road safety management activities and enable safety practitioners at the DOT and other agencies to utilize the models’ results to allocate the limited resources to the sites with highest potential for crash reduction,” said Naraghi.

Additionally, to comply with Iowa DOT objectives to reduce fatal and serious injury crashes, the SPF severe crashes model was utilized to identify the top 10 segments statewide with the highest potential for fatal and serious injury crash reduction, diagnose their safety problems, implement the appropriate safety countermeasures, and evaluate the effect of the implemented safety measures.

Though the project has been implemented, Hans said work has continued to categorize secondary road segments and there may be potential for additional research and development of updated models. As examples, the data may be further specified by direction on divided roads, there may be potential to examine municipal street segments, and opportunities to investigate targeted crash types such as lane departures and wet weather.

More information about the project and its implementation is available from the Iowa DOT.