Preventing Accidents Using Artificial Intelligence
Driver error, including distracted driving, is far and away the biggest cause of car accidents in the United States. However, another significant factor, according to a study by the U.S. Department of Transportation in 2008, is vehicle failure. The study found that of the accidents that weren’t caused by driver error over a 2.5-year period, 30.8 percent were caused by vehicle failure.
The most common vehicle failure leading to an accident was tire or wheel failure or tire degradation (43.3 percent of vehicle failure accidents) followed by brake failure (25 percent). These issues are easy to prevent if you know what to look for, but most of us don’t know the warning signs or have so many other everyday worries that we don’t remember to check for them.
UVeye’s automated scanning devices and deep learning algorithms allow you to monitor your vehicle and be alerted of any potential warning signs without having to recognize them yourself. UVeye allows you to carry out predictive maintenance to prevent an accident due to vehicle failure.
Artemis
Since tire failure or degradation is the most likely vehicle issue to cause an accident, it seems important to be able to check your tires and ensure that they are safe to drive on, with no signs of imminent failure. The current standard method for tire inspection is to take a coin and stick it between the treads to check for signs of routine wear on the tire. We think that in 2020, and with so much riding on the well-being of your tires, we can do better. Artemis, a product from UVeye for tire inspection, captures sharp images from passing vehicles and can identify any issues or warning signs in the tires.
Helios
Meanwhile, Helios is another product offered by UVeye that takes high-quality images of the undercarriage of a vehicle to identify any anomalies, damage, or warning signs. Brake failure is another common cause of accidents, and Helios can identify potential brake issues in advance, preventing future accidents. These issues may be barely visible to the human eye, but Helios can detect them within seconds.
Deep Learning
UVeye has put a ton of work into teaching the algorithms about the different parts of the car and how they look on different makes and models, and in different situations. The appearance of a car and its parts will change over time, and it can even change on a daily basis depending on weather, conditions, and other factors. A car that has just driven through dirt or snow will look different than it would under normal conditions. Our algorithm needs to be able to recognize parts from many different vehicles in all of these different conditions.
A New Standard for Vehicle Inspection
We hope to create a new standard in vehicle inspection and to catch issues early on that otherwise could have ended up causing an accident on the road. We can’t stop all auto accidents, but even reducing the number by a small percentage could save thousands of lives.
