Google’s Machine Learning Can Help Drivers Avoid Traffic, Damage From Potholes

If you’re the type of person who enjoys going out for a drive in places with beautiful scenery such as Southern California, then you probably know just how bad traffic can get.

Based on a report by INRIX, Los Angeles has the worst traffic in the world, with a record of 102 hours of congestion during peak hours in 2017. It’s why two Loyola Marymount University students decided to use machine learning in order to better understand traffic.

This meant analyzing elements such as road damage from potholes and cracks, with the help of TensorFlow, Google’s open-source machine learning platform. The goal is to train a model that can identify potholes and other road imperfections quickly using camera footage.

According to the students, both construction companies as well as cities could use such technology to identify which specific roads are in need of maintenance, which means that they would be spending more time fixing the problem instead of scrambling and trying to first identify it.

In a future where such measures are implemented, safer driving conditions and more efficient roadworks should lead to a better flow of traffic, not to mention happier drivers. Not just because of reduced traffic, but also because they would be spending less money on gas and fixing damages caused by poorly maintained roads on their vehicles.


  • MJack

    Pro: now you can slalom anywhere. Con: your passengers will hate you

    • JGreen

      I have the potholes memorized and slalom around them every day. I’m surprised I haven’t gotten pulled over on suspicion of drunk driving.

  • TheAmerican

    Or… you can look ahead and pay attention to the things around you. Problem fixed! No money needed for that one. Aren’t I smart?

  • ksegg

    We are headed for a Wall-E future.

  • Dylan Wentworth

    Fixing roads won’t be necessary for long. Once the auto makers have removed all the cars from their product lines and everyone’s driving a big truck, potholes are no big deal. Just install some 35″ tires and you won’t even notice them.

Prancing Horse Invites The Public To Maranello With Universo Ferrari Event

The Universo Ferrari event will allow visitors to see the company’s best road and race cars of all time.

2020 Mercedes-AMG GLC 43 4Matic Launches With 385 HP

The new 2020 Mercedes-AMG GLC 43 4Matic features a twin-turbo 3.0-liter V6.

2020 Cadillac CT5 Costs Nearly $10,000 Less Than The CTS

Tired of fighting the Germans head-on, Cadillac’s CT5 looks like a bargain.

Tuned BMW 640i Doesn’t Give A Hoot About The New 840i

Senner has made this 640i Convertible more powerful than its 8-Series counterpart.

Continental Says Fully Autonomous Cars May Not Arrive Before 2030

Tech needs to be improved, regulation put in place and the public must trust them – and that will take time.

New Lotus Sports Car Coming Late Next Year, Will Be Launched In 2021

The upcoming model could be a replacement for the Elise, Exige or Evora.

Alpina Once Considered Making A High-End Mini

Alpina couldn’t justify the complexity and cost involved in making its own version of the Mini.

Ford Unveils Their Smallest Ever Pickup, Could Be Launched Early Next Year… As An Emoji

Boasting a 0-60 mph time of never, Ford’s latest pickup could be headed to a smartphone near you.

2020 BMW 1-Series: New A-Class Rival Shows Us All Its Secrets

A huge, 179-image gallery shows BMW’s brand new entry level model in every detail.

Some 2019 Subaru Legacy And Outback Owners Could Get A Brand New Car – Here’s Why

Improper welding by robots may render the issue unfixable, in which case the cars will be replaced .