A team of engineering researchers at the University of Waterloo has developed new algorithms that can detect when people participate in distracting activities such as texting while driving.
The newly developed system uses artificial intelligence and cameras to detect “abnormal” hand movements; movements that indicate that a driver is participating in distracting activities and putting himself/herself and others at danger.
The information that the system provides can be used to improve road safety by alerting drivers when they’re dangerously distracted. And indeed, signs of serious driver distraction will be particularly useful in not only conventional cars but self-driving vehicles too.
“The car could take over driving if there were imminent danger, even for a short while, to avoid crashes,” said Fakhri Karray, an electrical and computer engineering professor at Waterloo.
The algorithms that the team used were trained using machine-learning techniques. Thanks to them, the algorithms can recognize actions such as texting, but also talking on a phone or reaching into the backseat to retrieve something, etc. The algorithms can tell when an action is super-serious and when it’s not based on duration and other factors.
University of Waterloo via ScienceDaily (https://www.sciencedaily.com/releases/2017/09/170908205543.htm)