The George Washington University Center for Intelligent Systems Research (CISR) is developing a new, inconspicuous drowsy-driver detection system.
Using artificial neural networks – which attempt to mimic the brain’s neuron processing for such functions as pattern recognition, signal processing, and control systems – CISR researchers are able to track and classify steering behavior, the college said. The system can detect the differences between drowsy driver behaviors vs. non-drowsy with 90% accuracy, researchers said.
CISR’s drowsy driver experiments in the simulators have revealed drivers’ steering behavior deteriorates for up to 2-3 minutes before an eminent crash. This dangerous behavior can be detected by CISR’s smart signal processing system well in advance of a potential accident and help motorists avoid hazardous drowsy driving.
Driver fatigue is the cause of an estimated 100,000 crashes annually resulting in more than 40,000 injuries, according to the National Highway Transportation Safety Administration. The agency’s Fatality Analysis Reporting System (FARS) indicates an annual average of 1,544 fatalities due to driver drowsiness related accidents, with fatigue involved in 10% to 40% of crashes on long motorways and 15% of fatal single vehicle truck crashes.
