U-M Technology Could Detect Impaired Drivers Using In-Vehicle Cameras

  • University of Michigan engineers have developed technology that could detect impaired drivers using cameras
  • The technology analyzes blood flow, heart rate, eye behavior, body posture, head position, and respiratory rate from a 20-30 second video of the driver’s face
  • It uses artificial intelligence to look for signs that intuitively indicate impairment like increased heart rate or unusual eye behavior
  • The system performs “anomaly occurrence detection” to flag abnormal behaviors compared to a baseline
  • The technology could be added to existing Advanced Driver Assistance Systems cameras in cars for a modest $5-10 cost
  • The researchers are seeking partnerships with automakers in Michigan to help address the ongoing problem of drunk driving