Computer algorithms developed by engineers from the University of Waterloo can accurately detect when drivers write SMS while driving or participate in other distractions. The system uses cameras and artificial intelligence to detect movements of hands that deviate from normal performance, classifying them in terms of possible security threats. According to experts, this development can be used to improve road safety by warning drivers that they are distracted.

The algorithms underlying the new technology use machine learning techniques to recognize actions such as typing text messages, talking on the mobile phone or turning towards the rear seat in order to get something, aiportal.ru reports. The danger of the action is assessed on the basis of the duration of its implementation and a number of other factors.

This work is based on extensive previous studies related to character recognition, which may indicate that the driver begins to fall asleep behind the wheel. The position of the head and face is also an important signal of distraction, experts say. Current research is aimed at combining the detection, processing and evaluation of several different types of distractions in one system.