Using AI to predict and prevent road accidents

by Startacus Admin

The lowdown on Predina, the startup using machine learning to make a positive social impact, using AI to dynamically predict and prevent road accidents in real-time...

In 2018, 26,610 people were killed or seriously injured on UK roads. Speeding, using phones, and simple reckless driving are just a few of the human errors that add to the dangers already posed by manoeuvring a hunk of metal around the roads at high speeds.

We tend not to think about it when driving, but the reality is that any number of small things could cause an accident, from animals running across the road, to an unexpectedly deep pothole, to a puncture, even to someone’s bin blown into the road by the wind - the dangers are sometimes small and mundane and things we would never think to actually call ‘dangers’.

Tech giants like Tesla are partially combating the dangers of driving with self-driving cars that can, in theory at least, predict an accident faster than we can and take action, and the likes of UK startup Wayve are tackling autonomous driving services. As with just about everything else, the solution seems to be artificial intelligence.

London-based machine learning startup, Predina, is taking on the bigger and more common issues of road dangers from outside the vehicle. The startup uses deep learning and machine learning models, which learn from more than 28 different factors - including weather, traffic, satellite imagery, and road layout - and over 2 million road accident records in order to predict the risk of a crash at a given place and time, and for any type of vehicle.

It can predict the likelihood of a crash, the likely severity and cause, and provide the safest route. And it does this up to 48 hours before you even leave the house. This could become a vital dynamic risk platform and has already shown a 70%-83% accuracy in prediction, 50% improvement in driver awareness, and 25%-53% reduction in road traffic crashes is their real-world commercial pilot.

Predina is designed to be implemented into any existing application, giving your customers greater safety and peace of mind as well as setting your product apart. It can also be implemented in the systems of a smart city, reducing congestion, cutting costs caused by accidents, and saving lives.

In March, Predina was named as part of the 6-strong 2019 residency cohort of Google for Startups, which will spotlight startups using machine learning technology to make a positive social impact. These 6 startups will startups will get access to Google’s products and connections, as well as tailored mentorship and workspace at the Google for Startups London campus.

We don’t need machine learning to predict good things for this startup!

Subscribe to our newsletter

If you would like to receive our startup themed newsletter, full of the latest startup opportunities, events, news, stories, tips and advice, then sign up here.

• #
• #ai
• #machine learning
• #social impact
• #mobility
• #traffic

Published on: 7th April 2019