University of Waterloo creates AI-based system for detecting distracted driving

Published: September 11, 2017

Updated: July 24, 2018

Author: Luke Jones



The University of Waterloo (UoW) has announced a new project that could prevent distracted driving activities in Ontario. Researchers have developed algorithms that can accurately detect when a driver is texting or using his smartphone while driving.

In an announcement on Friday, the university says engineers have devised a system that merges camera technology with artificial intelligence. The new system reads hand movements to determine whether a driver is operating a device. Hand movements differ between driving operations and other activates, the system can read the subtleties in the differences.

Using the information, the system grades the movements and levels them compared to safety threat.

In its press release, UoW says the new method can help to improve safety and alert drivers when they are distracted. It could even function with connected devices and shut down vehicles if the driver is distracted.

“The car could actually take over driving if there was imminent danger, even for a short while, in order to avoid crashes,” said Fakhri Karray, a university research chair and director of the Centre for Pattern Analysis and Machine Intelligence (CPAMI).

Researchers used machine-learning to train the artificial intelligence to recognize actions, such as texting, and applied it to the system’s algorithm.

CPAMI is continuing its research and wants to merge the detection, processing and grading system into a realized software. Karray says distracted driving is estimated to be the cause of 75% of all traffic collisions around the world.