Computer scientists at Queens College and Columbia University have developed software to identify if someone is drunk based on their speech. They believe someday the technology could be added to cars to keep people from driving while under the influence of alcohol.
The software is based on the work of researchers in Germany. Over a two-year period the researchers collected data from 162 participants in various states of drunkenness. They identified speech patterns often associated with intoxication, such as stammering, stuttering, and slurring.
The U.S. scientists believe their software could be used as a preventative measure to keep a car from starting if it determines a driver is impaired. However, they acknowledge the technology is several years away from practical use—in its current state the software correctly identifies a driver’s drunkenness or sobriety less than 75% of the time.
The drunken speech software is another component in the trend to use in-vehicle technology to outwit would-be drunk drivers. All 50 states have laws that require ignition interlock devices for some DUI offenders. And in 2015, the National Highway Traffic Safety Administration (NHTSA) starts the third phase of research on the Driver Alcohol Detection System for Safety (DADSS) in hopes of developing passive, non-invasive alcohol detection sensors that could be added to all cars.
While NHTSA statistics for 2013 show that the numbers are once again decreasing, annual deaths from alcohol-related crashes still top 10,000. Is creating technologies that allow vehicles to stop people from driving while intoxicated a step in the right direction?
