Skip to content

Khattak’s Research Could Help Cut Back on Big Traffic Incidents

Birds-eye view of congested traffic on an interstate at night.

For a commuter, nothing quite sinks the heart like having to slow to a stop in a seemingly never-ending line of traffic.

According to the 2019 Urban Mobility Report, in 2017, the average commuter sat in traffic for nearly seven working days, accounting for more than $1,000 in personal costs. Also alarming, commuting delays were 15 percent higher that year when compared with just five years before.

This trend also correlates with an increase in crash frequencies. Larger-scale traffic incidents in particular can come with challenges in the incident management process, especially if multiple agencies such as police, fire, and emergency medical response are needed. Gathering information about incidents is a chronological process, and it can be time consuming to collect all the necessary information needed for a tightly coordinated response.

Asad Khattak.

Asad Khattak

Department of Civil and Environmental Engineering Beaman Professor Asad Khattak recently collaborated on a study, published in Transportation Research Record: Journal of the Transportation Research Board, to develop a sequential prediction method to handle the chronological process of incident information gathering. The method is based upon parametric survival modeling, which is a statistical technique used to link the duration of an incident to covariates often utilized to predict incident duration.

Khattak said the study was motivated by a disproportionately high effect of large-scale incidents on traffic. He further elaborated that accurate and timely prediction of incident durations using advanced statistical and artificial intelligence tools will help us better manage large-scale incidents.

A wide range of studies exist regarding traffic incident duration, but few studies have attempted to analyze the incident duration by incorporating chronological incident information gathering.

This study took advantage of a unique incident database and identified more than 600 large-scale incidents in the East Tennessee area from 2015-16. From this data, the team was able to develop a five-stage prediction method according to the chronological process by which information becomes available during incident operations.

Future research is needed to dig deeper into various incident management areas such as on-site operational sequences, crash injury severity information integration (e.g., the reason the multi-agency response is requested), and traffic detour operations resulting from huge travel delay after large-scale incidents, all of which are based on additional information obtained from other sources.

An accurate and timely prediction of incident duration could help manage large-scale incidents in a proactive manner.

The flagship campus of the University of Tennessee System and partner in the Tennessee Transfer Pathway.

View our Privacy Policy.