AI can detect those who text and drive
In newer vehicles, automated safety features make getting behind the wheel safer than ever. Some cars now provide an alert when the vehicle drifts out of its lane or if another car is driving up in a blind spot. The services do a good job of making drivers more aware of their surroundings and forcing them to pay more attention.
But they are not perfect.
Author: Justin Tejada, The Connected Car
For one, it can be annoying to hear a beeping alert if you're just moving slightly out of your lane to allow a little extra room while passing a cyclist. The services also do not currently prevent one of the most dangerous behaviors: texting while driving.
But a new system is in development at the University of Waterloo in Ontario, Canada. At Waterloo Engineering, researchers are working on artificial intelligence-powered computer algorithms which, in conjunction with in-car cameras, can detect when drivers are engaging in distracted driving activities like texting.
The project comes out of the university's Centre for Pattern Analysis & Machine Intelligence (CPAMI), which conducts research in six areas: Machine perception, robotics and autonomous systems, cooperative intelligent systems, human-machine interaction, pattern recognition and image analysis, and data mining and knowledge discovery.
"CPAMI's findings will have the potential to address numerous real world problems that traditional systems cannot resolve," reads the Centre's website. "The envisaged application directions include, but are not limited to, public security, manufacturing, transportation, assistive environments for seniors and people with disabilities, communication, eLearning, finance and web services."
The driver distraction detection software incorporates previous research that centred around signs that a person may be at risk of falling asleep, including certain blinking patterns, positioning of the head, and facial expressions.
Other actions that are a little easier for the technology to detect, like texting or reaching into the backseat, are analysed based on duration to determine whether the driver needs to be warned.
For example, while briefly glancing at your cell phone may be ill-advised while driving, it likely wouldn't garner an alert from the system. Taking the time to send a text message, however, probably would.
As autonomous vehicles become a more integral part of our transportation systems and come to represent a greater percentage of cars on the road, the researchers see their safety system working in tandem with autonomous systems to help drivers avoid collisions.
"The car could actually take over driving if there was imminent danger, even for a short while, in order to avoid crashes," noted Fakhri Karray, director at CPAMI and a professor of electrical and computer engineering at Waterloo.
While plenty of media attention gets placed on the research and development of AV and AI safety systems done by automakers and software companies, it's nice to see similar work being done by an institution of higher learning, whose interest is more focused on the public good rather than the bottom line.
Additionally, the more groups that are working on vehicular safety -- and the greater the diversity of perspectives involved -- the faster we'll see more effective technology on the market.