Innovation: Driverless Cars
The real breakthrough in self-driving cars is inside the car…
By Dave Lashmet
biotech and technology expert
We’re in the early stages of an unstoppable trend that will transform your daily commute. You’ll never have to drive to work again…
You won’t have to deal with a never-ending parade of brake lights anymore. You’ll be able to finish the book you’ve been reading for weeks… catch up on dozens of e-mails before you get to the office… or help your kids finish their homework on the way to the bus stop.
And it’s a lot closer to happening than you might think…
For example, carmaker Audi just announced a new model A8 – with “Level 3” autonomy. Level 3 means hardly any attention is required to control the car. It basically drives on its own… And the driver doesn’t have to be ready to intervene in an emergency. When this model goes on sale next year, it’ll make history as the most highly automated car on the road.
Moreover, Audi is aiming for Level 4 vehicles by 2020. This means driver attention is never required.
The path to get to Level 4 depends on one company… And the real story in self-driving cars is what’s going on inside the car.
You see, self-driving cars originally needed to “memorize” both maps and how to use a car’s controls – how to accelerate, steer, and brake. But it would have taken programmers a long time – some estimates say up to 100 years – to teach a car every scenario.
Nobody was going to wait a century for this technology… Thankfully, one company has developed a breakthrough way for these self-driving cars to quickly learn on their own.
This company is best known for making graphics cards – a special kind of silicon chip that works alongside a central processing unit chip with the goal of creating images, especially in video games. The trick to this is a kind of computing called “parallel processing,” where lots of little math problems are all done at once… like coloring each portion of a digital picture.
But graphics cards are also necessary in industries far removed from video gaming… And what this company is doing with self-driving cars is the opposite of running a video game.
When you’re playing a video game, graphics cards take artificial objects and paint them into a visual scene. But this company figured out that if you can do this in a believable way in a game environment in real time, you can probably also do this in the real world… with real cars. This is the springboard for all the self-driving cars you’ll see in the coming years.
In a self-driving car, this company’s chips take real objects and paint them into a “game.” They use what they have learned in games to follow real streets and track real pedestrians…
Why Self-Driving Cars Work Like Video Games
Industry lawyers like to call developments in the field of self-driving cars “advanced driver-assistance technologies” to avoid liability for life-or-death decisions.
For example, “driver assistance” technology that applies the brakes or pulls you back into your lane is much easier to defend than technology that leaps curbs to avoid a head-on collision – but which in turn crashes the car into a school bus stop.
The key will be in absolving carmakers from liability. That’s when we will see truly driverless cars. Most pundits think that this is still five to 10 years out.
Self-driving cars aren’t good enough yet, largely because of weather and unexpected conditions that really can’t be programmed for: like a truck losing its load or a tree falling across a road. Driving on a bright, sunny day is easier than when it’s icy with blowing snow. Driving conditions deteriorate and create more risk, with no one willing to take the blame.
Fortunately, one company is tackling the first steps to solve these problems… We’re talking about California-based Nvidia (NVDA), a technology titan.
The trick to Nvidia’s strategy is that it’s like playing a video game in reverse. Instead of creating game-world objects rendered for the player’s point of view, Nvidia is tracking real-world objects to predict their motion from the perspective of a driver’s moving car.
Imagine that you’re tracking a semitrailer… Odds are that it weighs between eight and 28 tons and has good tire tread. So in the next five seconds, it can go five miles per hour faster if it steps on the accelerator… or 30 miles per hour slower if the driver mashes the brakes.
Nvidia has calculated an expected tracking range for cars, delivery vans, semitrailers, bulldozers, motorcycles, bicycles, and pedestrians walking beside and across a road… even dogs and cats. So it can predict future motion.
We saw this technology running in the Nvidia booth a year ago at the 2016 Consumer Electronics Show (“CES”) in Las Vegas. We walked in front of a car at a crosswalk, and we could see our speed and trajectory uploaded into a “game” that a smart car was playing.
So just by integrating fixed objects like streets, curbs, and stop signs with the moving-objects data, a self-driving car can reasonably navigate a cityscape. Nvidia is taking lessons learned in video games and applying them to real life…
Obviously, we were impressed. But Nvidia’s latest technology is going a step further. We believe its technology will become the future of the Internet and mobility. Technically this process is called “deep learning,” but it’s easier to think of it as “self learning”…
This Car Learns Like a Student Driver
The next step for a car is to learn to drive from “scratch”… without a programmed database of every curb, stop sign, and left-turn lane stored in a memory bank. All it uses is a map…
For months, Nvidia’s car “watched” data of what real-life drivers did in reaction to different scenarios. Then Nvidia took the car out to a closed course and let it start hitting cones. After a while, it started hitting fewer cones… and then no cones.
From there, the research team let it drive around cemeteries with no street names and no curbs. Then the car went into suburban streets. And finally, it started driving along the New Jersey Turnpike.
This is also the logic that will help a car negotiate snow, ice, and slush as proficiently as any human driver. That’s because the car can do more than just learn from its own experiences… It can also upload data from other cars.
Now, think back to our example of liability for a self-driving car. No command lines exist that order a car to leap the curb to avoid a head-on collision, but which regrettably wipe out a bus stop. Because there’s no command code, there’s much less liability.
In summary, self-thinking cars can out-think a fixed-object program that’s based on measuring every curb. The latest self-thinking car can also handle inclement weather. And it reduces carmakers’ liability.
It’s also 1,000 times cheaper to put in a $1,000 graphics card for a self-thinking car, rather than a million-dollar computer system with its massive database of pictures the car has to match. That’s how graphics cards pulling away from conventional computers pays off.
Auto manufacturers built 90 million new cars and trucks worldwide in 2015, according to industry group International Organization of Motor Vehicle Manufacturers. So at $1,000 for an in-car graphics card, each 1% market share is worth $900 million to Nvidia.
Machines are learning to think for themselves. This has never happened before. And the world might be forever changed. The upside is that early investors will profit the most…
Plus, at this year’s CES in Las Vegas, we watched as Nvidia unveiled its next generation of chips… a technology called Volta. Using the newest Nvidia chips, an Audi at the CES learned to drive around a maze – a closed course – in just four days.
This new electronics board wasn’t a one-off, either. Nvidia is sampling it to select customers now, and German car-parts manufacturer Bosch is preparing to buy tens of thousands of Volta chips.
That’s because carmaker Audi is going to put the prototype we just saw at the CES in full production for their 2020 model. Given the strange calendar for cars, that means we’ll start to see these on the roads in 2019.
And if you know anything about an automotive production line, you know that every component has to be finished and tested before the car gets assembled. So anything going in the 2020 model year has to be fully prepared by early 2019… And the chips have to appear in bulk by 2018.
In program management, that’s called a “work-back schedule.” You can think of it like preparing for a camping trip: You have to buy rain gear before you can pack it. Cars and chips are the same way. If you want them by a certain date, you have to subtract out the production time.
Seeing the Volta chip in the hands of Nvidia CEO Jen-Hsun Huang at this year’s CES and knowing that both Bosch and Audi are relying on this part, we know it’s in sampling now.
After sampling comes low-rate production, then high-rate production. That carries us into 2020… and a self-driving car. This Volta chip is the size of a postage stamp. And our risk was that it would never be built. Now, we’ve seen it… and seen it working.
After we saw Huang’s presentation and heard from Nvidia’s automotive vice president at this year’s CES, we met the company’s senior director of automotive technology and saw the car drive on a track.
The car – without a driver – could drive safely through construction zones, sand, dirt, and messed-up lane lines painted on the road… all with one camera. We know that as you add more and more cameras – and radar – to a car, it gets smarter. Plus, if you match the camera feed and radar to a map, accuracy improves even more.
That’s what Audi will build into its 2020 cars. But Audi has been making cars for 115 years. The new part is the car without the driver… and that’s from Nvidia.
Even if you aren’t sold on the benefits of a car that drives itself, this technology is also a highly advanced cruise control to deal with abrupt wind gusts or a fog bank that human eyes cannot see through. Plus, all these smart cars will be connected using mobile-phone technology. So if one car hits an ice patch, all the other cars will know about it.
In the end, Nvidia’s technology will make us all better drivers and make roads safer. Which is an incredible accomplishment from a video-game company.
What’s This Worth to Nvidia?
Nvidia trades like a mature tech-growth company…
It has a well-developed graphics-card business and invests in new markets… A great analogy would be Alphabet (GOOGL) with its search business that makes the money and its side projects – like driverless cars – that provide upside opportunities.
That’s good company to be in.
For Nvidia, its steady growth comes from the gaming sector… And its breakout opportunity is a brand-new market in automotive driver-assistance technologies, ultimately yielding a self-taught, self-driving car.
Here’s the implication: When machines begin learning for themselves, it will be a watershed moment in human history. And as we said, these “artificial brains” will bring huge profits to early investors.
Case in point, texting while driving… It’s now a leading cause of car crashes, right up there with drunk driving. It leads to more minor accidents and about one-third as many deaths as drunk driving, according to the Centers for Disease Control and Prevention.
Technology can protect us from that.
It’s hard to measure what this might be worth, but Huang believes self-driving cars will hit the market this year, much sooner than outside pundits think.
Practically, we agree. At first, we expect this self-driving car will be styled as offering “advanced driver-assistance technologies.” You’ll be able take your hands and your eyes off the road, and the car will take over… But you’ll own the liability.
This new market can add billions of dollars in revenue to Nvidia, starting later this year. Each 1% of the total vehicle market share… at $1,000 per graphics card… is worth $900 million to Nvidia. And although Nvidia is a well-known tech company, nobody on Wall Street is betting on artificial brains as a business segment.
We’re getting closer and closer to a point when you won’t even have to drive yourself to work. You’ll just sit back, relax, and let the car take you where you need to go. And Nvidia is playing a key role in the development of this technology. As we said earlier, investors who get in now will benefit most from the company’s massive future growth potential…
Dave Lashmet has spent 10 years teaching and writing about medicine and technology at major research universities, and he’s done follow-up research at some of the most important facilities in North America. He is also an inventor on three issued U.S. patents (in high-tech hardware and software), and a co-inventor of three more patent applications currently under review by the U.S. Patent and Trademark Office.
Dave writes Venture Technology, an exclusive letter that takes a “venture capitalist” approach to investing… seeking out small-cap speculative stocks with strong catalysts and outstanding breakout growth potential. His subscribers there are up as much as 185% in Nvidia.