Follow
Subscribe

Parallel Domain Looks to Train Autonomous Vehicles in Virtual Reality

Home > News > Content

【Summary】A former Apple and Pixar employee is developing technology to help companies and automakers test their self-driving cars in a virtual world.

Original Vineeth Joel Patel    May 04, 2018 12:06 PM PT
Parallel Domain Looks to Train Autonomous Vehicles in Virtual Reality
author: Vineeth Joel Patel   

Testing autonomous vehicles can be a tricky thing. Some companies, like Uber, have created fake cities with obstacles to test their self-driving vehicles. Others have gone with a more direct route by testing driverless cars in numerous cities, like Waymo has.

Virtual Testing Is Becoming Mainstream

Others have resorted to testing machines in virtual reality, as it's the safer alternative. The option also allows companies to create unique challenges that they can test over and over again numerous times. With virtual-reality testing become more mainstream, Parallel Domain, a provider of 3D environment generation software, is looking to take the burden off of companies by developing its own virtual world.

Founder and CEO of Parallel Domain, Kevin McNamara, was a former Apple and Pixar employee. At Apple, he reportedly had a hand at working on the company's Special Projects Group that specialized in self-driving vehicle simulation. The company has made its public debut with roughly $2.5 million in funding, which is from various partners, including Costanoa Venture and Ubiquity Ventures. With that money, Parallel Domain is looking to simplify virtual reality testing for autonomous vehicles.

"What we do is use computer graphics to try to accelerate the development of safe autonomous vehicles," McNamara told Tech Crunch. "The idea being that, in a simulation, you can safely make a mistake and then learn from those mistakes. In a virtual world, you're not going to hurt anyone in this simulation."

Currently, companies and automakers spend a lot of money and hours developing virtual reality simulators of its own. Parallel Domain wants to simplify things by offering its pre-built program to those that want to rack up miles safely and virtually.

How Parallel Domain's Tech Will Help Companies

The company uses real-word map data to construct "a photorealistic, living world complete with traffic, pedestrians, time of day, and more," said Parallel Domain. The company went on to state that anything in its virtual world is programmable, including the number of lanes and the condition of the road itself.

"Road curvatures, locations of mountains, and hundreds of new cities can be generated with a few simple clicks," claimed the company. Of course, things like the amount of traffic, time of day, and number of pedestrians can be adjusted, as well.

Nio, a Chinese startup that's developing electric and autonomous vehicles, will be Parallel Domain's first customer. While Nio is predominantly in China at the moment, the brand has plans to sell its vehicles in the U.S. by 2020. Those cars are expected to have some semi-autonomous capabilities. As the company outlines in a press release, Parallel Domain will help Nio and other brands to eliminate concerns and ensure that everything is working before putting cars on public roads.

There's no replacement for real-world testing, but Parallel Domain's virtual world seems like the next best thing. Waymo, which is at the top of the totem pole for autonomous testing, stated that it covered 5 million miles earlier this March. While that seems like it would be enough, researchers believe that self-driving cars will have to travel roughly 11 billion miles to get a good picture of how driverless vehicles handle real-world conditions.

"Driverless cars need massive quantities of challenging training miles in order to learn how to drive safely, but these real-world miles are dangerous expensive, and inflexible," said McNamara.  "State of the art simulations alleviate these bottlenecks while providing essential interactivity, control, and reproducibility."

Prev                  Next
Writer's other posts
Comments:
    Related Content