Machine Learning Revolutionizes the Battle Against Distracted Driving

Distracted Driving Study

Distracted driving is one of the top contributing factors to traffic accidents. According to the NHTSA, 80 percent of vehicular collisions are the result of some form of distracted driving. More recently, a company called Zendrive has been collecting data around distracted driving using the sensors on peoples’ smartphones. With machine-learning algorithms, Zendrive can collect data regarding aggressive and distracted driving and transform large data sets into “actionable safety insights about driving behavior for an individual, fleet, or road.”

Test Results

Already, Zendrive has completed a test involving 3 million drivers covering 570 million trips and 5.6 billion miles. The results? A lot of people are using their phones while driving. In fact, drivers use their phones 88 percent of the time.  And while on the road, phones take up 3.5 minutes of every hour. The study – the largest of its kind – also discovered that moving your eyes away from the road in front of you for two seconds can greatly increase your chances of getting into an accident.

Deep Learning

So what exactly do we do with all this information? Scientists at the University of Waterloo’s Centre for Pattern Analysis and Machine Intelligence have suggested one possible answer: taking deep neural networks and applying them to cars, so that they can tell when a driver is involved in risky driving behavior. The concept of deep neural networks, also known as end-to-end deep learning, was developed by Jeff Dean 25 years ago at the University of Minnesota. At the time of its creation, there wasn’t nearly enough computing power to effectively implement the technology.

The Expansion of Deep Learning

Now, years later, Dean, a rock star in Silicon Valley, works at Google where there are thousands of processors supporting the company’s gigantic network. With such an insane amount of computing power, deep neural networks have become more effective, being used by Google to take online language-translation to new levels. DeepMind, a lab in London, developed AI capable of beating a world championship (Lee Sedol) at the ancient game of AlphaGo – a feat no one thought possible, at least for another decade or so.


So what are deep neural networks and how do they apply to cases of distracted driving? End-to-end deep learning is a type of artificial intelligence that takes the human brain as its model. Though it’s not a veritable replica of the human brain, it’s designed to mimic the central nervous system as we understand it. As Dean told Wired Magazine, “If you have lots and lots of these neurons and they’re all trained to pick up on different types of patterns, and there are other neurons that pick up on patterns that those neurons themselves have built on, you can build very complicated functions and systems that do pretty interesting things.”

Smarter Cars

To apply this technology to the distracted driving problem, the University of Waterloo team, led by Fakhri Karray, utilizes a mixture of specialized cameras, known as Microsoft Kinect cameras, and dashboard cams. The scientists begin by entering a large number of possible scenarios into the computer – this includes various hand positions, head placement and other body images associated with distracted driving. Once the AI has a foundation of images to work with, it can then begin the process of “learning,” extrapolating from those images other scenarios not included in the original input. As Karray put it in an interview with Wired Magazine, “Unlike pattern recognition-based algorithms, deep neural networks learn from the huge number of samples presented to them to build their capabilities.”

Eventually, cars will be able to accurately determine when risky behavior is underway, setting off alarms to alert the driver of danger. And further down the line, cars could even take control of driving. The ultimate goal, according to Karray? Self-aware cars. With this kind of technology, it’s easier to imagine a world free from most traffic accidents.

If you or someone you love has been injured because of a distracted driver, contact an accident attorney in your area now.

Sean Lally About Sean Lally

Sean Lally holds a BA in Philosophy from Temple University where he also studied theatre for several years. Between 2007 and 2017, he worked as a professional actor for several regional theater companies in Philadelphia, including the Arden Theatre Co., EgoPo Productions, Lantern Theater and the Bearded Ladies. In 2010, Sean co-founded Found Theater Company, an avant-garde artist collective with whom he first started to cultivate an identity as a writer.