Startup automotive companies that want to develop self-driving technology often face a lack of real data. In this Backstage pass excerpt from “The AI/ML Show” recorded on January 12Motley Fool contributor Jose Najarro explains how Nvidiait is (NASDAQ: NVDA) Ominverse provides the advanced tools these companies need to power their self-driving cars.
Jose Najarro: So, now I want to talk a bit about Nvidia’s software. First, NVIDIA DRIVE Simulation. This is a digital world created on Nvidia’s Omniverse. The most important thing about this digital world is that it follows the rules of physics, and that’s what you need when collecting data, especially in the self-driving vehicle. You must know, if the wind is blowing in this direction, how should my sensors react? If there is some form of weather or motion, say rain is falling, hail is falling, what physics does that change with my overall vehicle? This allows the collection of artificial data to train your autonomous vehicle algorithm.
It also allows validation of events that occur infrequently or cannot be tested often. For example, there may be events where what should you do if a truck flips through the air? You can’t really collect too much data from this in real life events. If an event like this happens and you have a vehicle that has never seen this, you won’t know what that vehicle would do in this scenario. Having this digital world allows you to test simulations or events that might rarely occur and be able to validate them.
One of the examples might be the way Toby was talking about schools, how when those lights go on and off, if they were run in some form of online simulation, you would be able to see what’s going on during these events. I believe this digital world created with Nvidia’s Omniverse is a game-changer for late starters. One thing Trevor mentioned You’re here has a lot of actual mileage data among its competitors. I believe that to some extent this simulation event allows latecomers not to worry about how some competitors might have a lot of data. Because if they obviously invest a lot of money in servers, in these software solutions, they will be able to recover a nice amount of mile data or artificial data to compete with some of the leaders.
Then they also have what they called NVIDIA DRIVE Constellation. As I mentioned, Nvidia has this digital world. But what if you don’t have cars? Let’s say I’m a start-up and I don’t have any vehicles and I have no way to access the data. There is nothing wrong with that. You can buy two servers, which Nvidia calls NVIDIA DRIVE Constellation. One of the servers can run the digital world, while the other can run a digital car. Now you’re collecting real-world data without even having a vehicle. As I mentioned, it’s amazing for any start-up or for any company that doesn’t want to try the hardware just yet. They want to see how to program the algorithm before they even try to hit the road.
Next, I want to take a look at some of the software solutions. Nvidia has NVIDIA DRIVE SDK. For those unfamiliar with the SDK, that pretty much means a software development kit. Think of it as a toolbox with different tools to use. Every tool has a purpose. Nvidia has created many and many tools that are in this toolbox. It was the slide. This is the SDK they provided. You start with your SDK, your OS, your DriveWorks, and then you have different tools. Let’s go with all these tools.
First, the SDK, the development kit. This is the material we were talking about earlier. These solutions, the Hyperion-I which includes all the sensors, all the hardware you need. Let me go back to where I am, so that’s where you start. The previously mentioned hardware is these development kits. Then you have your drive operating system. It is the operating system. Think of it as your Windows, the foundation you need to use your computer. Nvidia has created its own operating system on which you can use these tools. Then you have NVIDIA DriveWorks. This is the ability to transfer your information from your sensors into your program. You have all these sensors, but if there’s no way to communicate with your software, there’s no point in having all these sensors. That’s what DriveWorks does. Then you have DRIVE AV and IX from Nvidia, they are all different tools. Here, let’s say you want to do something like AI assistance, co-piloting, let’s say you want to do something with mapping, with planning or perceptions, Nvidia has done all of these tools. I think it’s very important. I think this is my last slide.
I think this toolkit that Nvidia has created is very important because the only way for you to get these tools is by using Nvidia’s software solutions. This is how Nvidia gave its graphics cards the ability to have a purpose. Now, with more solutions they release in software, more automakers are going to come out to use more of their solutions. It’s more like a software subscription service and also a hardware subscription service.
This article represents the opinion of the author, who may disagree with the “official” recommendation position of a high-end advice service Motley Fool. We are heterogeneous! Challenging an investing thesis — even one of our own — helps us all think critically about investing and make decisions that help us become smarter, happier, and wealthier.