Smarter systems, faster response, safer roads

When a crash happens on a busy roadway, the clock starts immediately. First responders race to the scene, often with limited information about what they’ll encounter, from blocked lanes, poor weather conditions, or secondary hazards from ongoing traffic.

For both the drivers involved in the incident and the safety of the responders themselves, those first moments are critical; but, despite the urgency, many traffic incident management systems still rely on fragmented data and delayed reporting, making it harder to respond quickly and coordinate effectively.

As part of a newly funded US Department of Transportation Safe Streets for All initiative in Westmoreland County, CMU’s Mobility Data Analytics Center (MAC) is developing an AI-powered dashboard to provide real-time clarity in fast-paced situations. The system will integrate data from multiple sources—including crash reports, weather conditions, and live traffic feeds from platforms like 511PA, Waze, and INRIX—to predict road incidents in advance and give responders a proactive and complete picture of an incident as it unfolds.

“The system can be a great addition to real-time traffic operation to improve roadway safety and mobility,” said Sean Qian, professor in the Department of Civil and Environmental Engineering and Heinz College of Information Systems and Public Policy.

Rather than relying on one source of information, the dashboard combines these inputs to detect and communicate incidents faster, demonstrating the potential to notify first responders 15 to 25 minutes earlier than the current legacy system. As a result, this information can also reduce secondary crashes, ease congestion, and create an overall safer situation for the front-line workers, those involved in the crash, and bystanders.

A key component of the initiative is also building a more complete understanding of roadway safety by tracking near-miss events and responder injuries – data that is often overlooked, but critical for improving long-term outcomes. By identifying patterns across incidents, the system can help agencies predict risks and refine their strategies over time.

Westmoreland County serves as an ideal testing ground for this approach. A suburban-rural freight hub where Route 30, I-70, and I-76 converge, the region experiences a perfect storm of road safety challenges, including high-speed traffic, heavy trucking activity, and rural road conditions.

Through partnerships with state and regional agencies, the project will also support hands-on training for first responders at PennSTART facilities. Users will learn to operate emerging transportation tools such as drones and driving simulators, enabling new technologies to be tested before deployment in real-world scenarios.

“We are very excited to work with Westmoreland County and PennSTART to pilot and test this real-time system,” said Qian.

For the researchers, the project represents a broader vision: using data and AI not just to understand infrastructure systems, but to actively improve how they function in real time. The goal is to eventually scale the model deployment across Pennsylvania, offering a blueprint for how communities can modernize traffic incident management and achieve USDOT’s vision of “safe streets for all.”

---

This project is a collaborative effort between Carnegie Mellon University, Westmoreland County, the Pennsylvania Turnpike Commission, the Pennsylvania Department of Transportation, the Southwestern Pennsylvania Commission, and the Regional Industrial Development Corporation.