The Iowa Department of Transportation (DOT) is clear that people shouldn’t text and drive, but that doesn’t mean that cellular phone applications aren’t able to aid its operations.
The crowd-sourced app Waze provides real-time traffic updates for drivers, both through hands-free data collection and active reporting of incidents by passengers or stopped drivers. But the app is also providing data for Iowa State University’s Institute for Transportation (InTrans) to help the Iowa DOT better respond to incidents across the state.
Because it provides up-to-the-minute input on incidents, the Waze data has the potential to give the Iowa DOT alerts faster.
But more work is needed to make the data useful for Iowa DOT’s Traffic Management Center (TMC) operators, so they can spend less time tracking down potentially false or duplicate traffic incident reports.
“Waze has been a valuable tool not just for drivers but for Iowa DOT’s Traffic Management Center as well. But the data can often have redundant events which must be refined and filtered. With these added features, it will be an even greater tool to help our operators respond more quickly and efficiently to incidents,” said Sinclair Stolle, Iowa DOT Traffic Management Systems Engineer.
Currently, Waze data comes in by individual reporters, or individual cell phone data, which means there are often duplicate reports and the possibility of errors in the reporting. The data also doesn’t make clear the extent of the incident, so for example, it’s not always clear whether a crash is contained in the shoulder or is stopping all traffic.
So, researchers at InTrans’ REACTOR Lab are working to cluster those reports to get a better picture of what’s happening on site. The clusters are grouped by traffic jams, traffic crashes and a combination of jams and crashes.
Clustering the data saves operators the time of checking those potential errors and the confidence to know the incidents they look at are confirmed.
ISU graduate student Mostafa Amin-Naseri has been leading the work, which started by quantifying the value of the data, with the guidance of Anuj Sharma, REACTOR Lab co-director and ISU associate professor. They are now in the second phase and working to create a clean data feed -- without the duplicates and with a clearer understanding of the extent of the incident -- that can be used by the TMC operators.
While work is ongoing to further enhance the usability of the data, it’s already providing some useful insights.
Further research will work to improve confidence in how the clusters are grouped. Researchers will also integrate the data from Waze with data from other sources.
The goal is for the enhanced feed to provide reliable real-time updates on traffic that could be displayed on 511 map with clusters tagged by their reliability and whether it’s a recurring incident.