Introduction to AI Tool for Road Surface Conditions
ALBANY, N.Y. (NEWS10) — There are some exciting new innovative projects unfolding inside of the xCITE Laboratory at the University at Albany’s ETEC facility. Carly Sutter, a graduate research assistant, has been using her expertise in mathematics and merged it with her growing interest in weather forecasting to build the next AI tool to help keep roads in New York state clearer and safer during severe weather events.
The Importance of the Project
"DOT has a really important job to be able to keep road travelers safe. So we are building this tool that can be used in addition to their current process. It’s another tool to make decision making a little easier, a little faster," Sutter explained.
Building the AI Tool
For the last four years, she has been building her very own machine learning tool. Sutter has been feeding her model curated information in the form of hundreds of images from DOT road cameras that have been sorted and labeled to help the model better understand road surface conditions given the weather data during the time the image was taken. The more examples it takes in, the better it’s able to predict road conditions when presented with future forecasts.
How the Tool Works
"You have to tell it the rules, essentially. You give it a picture, give it some weather data, and you say, ‘Under these conditions, it’s snow; it’s wet.’ Once that model is trained, the whole benefit is we have new information on the fly that we can feed it. Stuff that’s it’s never seen before and make its prediction. That’s where the efficiency comes into play," Sutter said.
Future Plans
Sutter explained, eventually, it will be able to take in brand new images in real time from more than 2,000 DOT cameras dotted throughout the state and predict conditions to help crews mobilize faster.
"It’s not replacing anybody. Just another piece of information. If we can make road surface better by using it, we should absolutely do that," Sutter added.
Funding and Support
The development of the tool is funded in-part by $2.4 million awarded by the Institute for Research on Trustworthy Artificial Intelligence. Sutter hopes by winter of 2025 her tool will be fully functional for NYS Department of Transportation to start using, paving the way for a more automated process that could potentially save a lot of time.
Current Process
"Currently, DOT manually checks camera images," Sutter explained. "They get a lot of information about road surface conditions. The benefit of this is it can look at camera images instantly."
Statement from NYS DOT
NYS DOT said the following about the project in a statement sent to News10ABC:
"The State Department of Transportation regularly collaborates with institutions of higher learning on the latest innovations in engineering and highway safety, and this collaboration with UAlbany has great potential to help our Department better assess road conditions in real time."
Scott Cook, Information Specialist NYS DOT
Upcoming Presentation
On Wednesday April 30, Sutter will be presenting her research and showcasing the new tool at UAlbany’s Showcase Day, which hopes to bring together the entire school community to highlight innovative research.
Conclusion
The development of this AI tool is a significant step forward in improving road safety during severe weather events. With its ability to predict road conditions in real-time, it has the potential to save lives and reduce accidents. The collaboration between UAlbany and NYS DOT is a great example of how innovation and technology can be used to solve real-world problems.
FAQs
Q: What is the purpose of the AI tool being developed by Carly Sutter?
A: The AI tool is being developed to help predict road surface conditions in real-time, allowing for faster and more efficient decision-making by the NYS Department of Transportation.
Q: How is the tool being funded?
A: The development of the tool is funded in-part by $2.4 million awarded by the Institute for Research on Trustworthy Artificial Intelligence.
Q: When is the tool expected to be fully functional?
A: The tool is expected to be fully functional by winter of 2025.
Q: How many DOT cameras will the tool be able to access?
A: The tool will be able to access more than 2,000 DOT cameras dotted throughout the state.
Q: What is the current process used by NYS DOT to assess road conditions?
A: Currently, NYS DOT manually checks camera images to assess road conditions.