Pranamesh Chakraborty is researching traffic incident detection at the Institute for Transportation (InTrans), so he is well aware of the many advantages and some shortcomings of the different data sources that provide useful information in determining what’s happening on roadways.
Armed with that knowledge, Chakraborty was interested in developing a model that would fuse different information sources and build a reliable incident detection model.
His paper on that effort earned him the best paper award at the Transportation Research Board (TRB) sponsored 2018 Managing Roadways and Transit Meeting and Conference.
“It feels awesome to get rewarded for the work done. It also gives the motivation and the encouragement to carry out the research further and implement them in the real world, helping people to provide safe transportation,” said Chakraborty, who accepted the award in Bellevue, WA, last month.
Chakraborty, who is working toward his doctorate at InTrans, wrote the paper with fellow graduate students in Iowa State University’s Department of Civil, Construction, and Environmental Engineering Bijan Vafaei and Shefang Wang. The paper was part of a class project for Advanced Topics in Transportation Engineering: Data Analysis, led by Chakraborty’s advisor Dr. Anuj Sharma.
“We are also grateful to the reviewers and my advisor Dr. Anuj Sharma who provided insightful suggestions to improve our work further,” Chakraborty added.
The paper titled “Decision Fusion for Freeway Traffic Incident Detection” showed the work of the students who created a model that could fuse data sources like navigation applications that work best during peak driving hours and those like radar sensor based traffic data that work best when traffic flow is impacted.
“Our work showed that incident detection rates can be improved significantly with such a decision fusion model,” Chakraborty said.
The award was one of two papers that were selected by the TRB Freeway Operations, Managed Lanes, and Transit Management Performance committees. The reviewers who selected the paper praised it as well-written and also particularly relevant as more data sources become available. The winners received a $1,500 award.
Chakraborty’s dissertation work is on traffic incident detection using large-scale traffic data and cameras, particularly incident detection models using computer vision techniques and big data analytics. He also won the Intelligent Transportation Systems (ITS) America student essay competition sponsored by the SouthWest Research Institute in 2016.