Login

Nvidia Launches Drive Map, A Mapping Platform For Autonomous Vehicles

Nvidia has combined DeepMap and data sourced from cars running Drive Hyperion.
Calendar-icon

By Sahil Gupta

clock-icon

1 mins read

Calendar-icon

Published on March 26, 2022

Follow us on

google-news-icon
Story

Highlights

  • Nvidia has combined the technology with DeepMap with its own data
  • Drive Map will be available to third parties who compete with Nvidia
  • Nvidia is also using its Omniverse simulation tool for the same

Nvidia has launched a mapping platform geared towards autonomous cars called Drive Map. The map coverage encompasses 482,800 km of roadway across North America, Asia and Europe which will be achieved by 2024. This was revealed by Nvidia's founder and CEO Jensen Huang at the GTC event. 

The platform has a clear intent - it wants autonomy to be ubiquitous and it is a space the world's most valuable semiconductor company has invested heavily in. This technology, Nvidia intends to open up to third parties, just not restrict it to its own platforms. 

Alongside this, Nvidia launched the next generation of Drive Hyperion which is its sensor and compute self driving stack which is used already by the likes of Mercedes, Volvo, JiDU, BYD and Lucid Motors. Autonomous vehicle companies like TuSimple, WeRide, Amazon backed Zoox and DeepRoute.AI are also using Hyperion. 

uq3ujjvk

The startup was founded by ex-Google, Apple and Baidu employees

Drive Map is based on the technology from DeepMap, a company Nvidia acquired last year. It enables centimetre level accuracy by combining DeepMap accurate survey mapping data with anonymous mapping data that has been crowdsourced from vehicles that leverage Hyperion platform. The mapping tool kit is a combination of camera, LiDAR, and radar to provide layers of redundancy needed for autonomy. 

All the data pulled from Nvidia customers is being uploaded in real time to the Nvidia cloud as the car drives. Then it is aggregated and loaded on Nvidia's omniverse platform which enables virtual collaboration in real time enabling accurate simulations. It is used to update the map so the vehicles can achieve proper localisation. This is how Nvidia scales the map and ensures incredible accuracy. 

There are automated Omniverse tools which enable content generation like the building of a detailed map and then is converted into a drivable simulation that be used with the Nvidia Drive Sim. 

Calendar-icon

Last Updated on March 26, 2022


Stay updated with automotive news and reviews right at your fingertips through carandbike.com's Google News

Related Articles

Latest News