Big Data-Based Smart System Helps Iowa DOT Manage Traffic Accidents

The Traffic Incident Management Enabled by Large-data Innovations system manages traffic when there is a crash, a stalled vehicle or bad weather.

by Grayson Schmidt, Ames Tribune, Iowa / April 3, 2017
Anuj Sharma – an associate professor of civil, construction and environmental engineering, a research scientist for the Institute for Transportation and the leader of the TIMELI project -- talks to transportation researchers in the REACTOR lab. Christopher Gannon/Iowa State University

(TNS) -- Iowa's roads and highways may soon be made much safer thanks to a new smart system being develop by a group of engineers at Iowa State University. The project — known as the Traffic Incident Management Enabled by Large-data Innovations system, or TIMELI — takes Iowa Department of Transportation data, sorts through it all and quickly identifies problems, resulting in earlier detection of incidents, such as crashes and weather problems, and better traffic management.

According to a news release from ISU, TIMELI is a smart system for managing traffic when there is a crash, a stalled vehicle or bad weather. But ISU Associate Professor of Civil, Construction, and Environmental Engineering Anuj Sharma said the project will hopefully improve upon previous attempts, and be able to simplify the large amount of data successfully.

"There is a lot of infrastructure deployed along the roads, but most of it used to manually detect if something is going wrong on the roadways," Sharma said. "People have tried to automate all of these things, but at that point technology wasn't very reliable."

TIMELI allows researchers to call up all that data on the six big screens arranged around the Realtime Analytics of Transportation data lab, or REACTOR, at ISU, and filter out any oddity or data that stands out. Sharma said the main problem with previous technologies was the potential to crash during stressful situations. When dealing with vehicle accidents, and severe weather, Sharma said the technology needs to be reliable to make the job easier for those assigned to read the data.

"If you are in a very high-stress situation, and you even get 10 wrong calls or false alarms, that kind of frustrates the human users," Sharma said. "Once you get those things, you just shut down the technology and go completely manual."

According to ISU, advancements in machine learning will allow the TIMELI system to learn from experience and find ways to do a better job analyzing the Iowa DOT's data streams, finding incidents and maybe even predicting problems. To make the TIMELI system work, officials with ISU said researchers are working to develop new traffic models, computer algorithms, user-friendly computer displays and information visualizations that will help operators make decisions and take actions. They're also using big data technologies to design new systems for data handling, archiving, analysis and output.

In the news release, Institute for Transportation Associate Director Neal Hawkins said the system receives data every 20 seconds from high-definition camera feeds and sensor information every minute from all over the state.

"There is more data than you could ever imagine coming out of this system,"Hawkins said.

Sharma, who is also a research scientist for the Institute for Transportation and the leader of the TIMELI project, said the group is currently in the first year of three for the project, and hope to have a prototype developed by the end of the year.

"We are not aiming to come up with one solution and say that is the best solution," Sharma said. "We are trying to have a solution ready, and improve further and further as we get the info from agencies."

©2017 the Ames Tribune, Iowa Distributed by Tribune Content Agency, LLC.