Home Automotive Nvidia has a finger in many automotive pies

Nvidia has a finger in many automotive pies

While most IT people know Nvidia as a graphics card and GPU vendor, the company's products are increasingly being adopted by the automotive industry.

Nvidia used this week's GPU Technology Conference to showcase some of the work being done by the company and its customers in the automotive field.

The Nvidia Drive platform covers the spectrum from improved dashboard instrumentation to fully autonomous cars, explained Nvidia senior director of automotive Danny Shapiro.

Mercedes Benz is using Nvidia Drive to explore various ideas, including high-quality displays that reflect the design of the car (perhaps even adjusting to suit different lighting conditions) using the Drive CX digital cockpit computer (pictured above).

It has also developed a pedestrian detection system (running on the Drive PX autonomous car development platform) that uses cameras and lidar to work effectively under a variety of conditions including bright light, darkness, and rain.

Mercedes pedestrian detection

When tested against the Cityscapes Dataset, Mercedes Benz's system has achieved almost 90% accuracy for detecting cars, and also detects pedestrians, road signs and certain other objects.

Audi is another car manufacturer working with Nvidia's kit, and even though its vehicle recognition system was trained in good weather conditions, it still performs very well when visibility is poor, Shapiro noted.

He explained that Nvidia's underlying approach to autonomous cars is to collect data from one or more cars, which is then uploaded to the cloud and used to train a neural network that network is then loaded into autonomous vehicles, where it is able to make realtime decisions.

New data is collected by the vehicles as they operate, and fed back to the cloud so the network can be frequently retrained and the improved version used to update the cars. For example, if the initial training was in urban areas, the system would probably never see a kangaroo, so the network would initially just treat one as an obstacle if it was actually in the road. But over time the network might learn that a roadside object that looks like a kangaroo is relatively likely to move into the car's path, and therefore it should slow down.

That is part of the 'magic' of deep learning: the ability to build what are effectively complex rules based on experience. Human drivers develop a 'sixth sense' that another driver is about to change lanes dangerously close in front of them, or that a pedestrian is about to step off the footpath - even if they can't articulate the clues they are responding to.

As previously reported, Nvidia is backing the Roborace competition for autonomous electric racing cars. All 20 cars will be identical except for livery and software, so the contest is really about developing the best software to run on the Nvidia Drive PX 2 card that will serve as the brain of each car.

Autonomous race car

Shapiro added that one slot will be reserved for an 'AI community' team to make sure that good ideas from outside the nine other teams (which will presumably be drawn from large companies and research centres) are given an opportunity to prove their merit.

The Roborace car is expected to make its debut this northern summer, he said, and the plan is run at least two types of race, one where contact with another car is forbidden, the other where it is permissible.

The latter is presumably to make the event more exciting for the crowd, although in a keynote presentation not much more than an hour earlier Toyota Research Institute CEO Gill Pratt drew attention to the empathy shown by spectators towards robots 'injured' in competition. Perhaps that only happens when they are at least vaguely humanoid.

Disclosure: the writer attended the GPU Technology Conference as a guest of Nvidia


Australia is a cyber espionage hot spot.

As we automate, script and move to the cloud, more and more businesses are reliant on infrastructure that has the high potential to be exposed to risk.

It only takes one awry email to expose an accounts’ payable process, and for cyber attackers to cost a business thousands of dollars.

In the free white paper ‘6 Steps to Improve your Business Cyber Security’ you’ll learn some simple steps you should be taking to prevent devastating and malicious cyber attacks from destroying your business.

Cyber security can no longer be ignored, in this white paper you’ll learn:

· How does business security get breached?
· What can it cost to get it wrong?
· 6 actionable tips



Ransomware is a type of malware that blocks access to your files and systems until you pay a ransom.

The first example of ransomware happened on September 5, 2013, when Cryptolocker was unleashed.

It quickly affected many systems with hackers requiring users to pay money for the decryption keys.

Find out how one company used backup and cloud storage software to protect their company’s PCs and recovered all of their systems after a ransomware strike.


Stephen Withers

joomla visitors

Stephen Withers is one of Australia¹s most experienced IT journalists, having begun his career in the days of 8-bit 'microcomputers'. He covers the gamut from gadgets to enterprise systems. In previous lives he has been an academic, a systems programmer, an IT support manager, and an online services manager. Stephen holds an honours degree in Management Sciences and a PhD in Industrial and Business Studies.


Popular News