Skip to main content

A Topology-Aware Incident Detection System and Interactive Dashboard for Highway Incidents

Invention Reference Number


Licensing Contact

Andreana C Leskovjan
(865) 341-0433
Traffic on a Detroit highway

Highway incidents including severe accidents can result in injuries and fatalities, and rapid emergency response is essential to preventing loss of life. Additionally, a great deal of time is lost due to traffic jams that result from highway incidents. Sensors on major highways can detect when an accident occurs, however it can often take several minutes for emergency services to be called. Estimates are that 10,000 lives could be saved each year if emergency response was just one minute faster. This technology detects highway incidents through machine learning and displays vital information on an interactive dashboard, and can decrease detection time by 4 minutes. 


This technology is a quick and reliable automatic accident detection system for highways. This system provides a way to significantly reduce detection time of a highway incident, and still maintain high operational performance. Existing traffic sensors provide information on speed, volume, and lane occupancy, which are constantly being tracked. But the timeframe they work on is not always useful. This technology uses machine learning radar-based traffic sensors. The technology uses more than just one radar sensor, as current technologies rely on a single point of reference. By utilizing data from neighboring sensors and creating a topological representation of the sensor network, this system can provide a tool for traffic operations professionals to quickly respond and to investigate before a 911 call is even made. This system can increase detection time by 4 minutes – from 5 minutes to 1 minute – increasing the opportunity to save lives. 

Applications and Industries

  • Highway operations centers 
  • State governments 
  • Traffic engineering industry 


  • Better way to detect accidents and incidents on highways 
  • Decreases incident detection by 4 minutes (from 5 minutes to 1 minute) 
  • Reduces time needed for emergency response 
  • Saves lives