NVIDIA Powering the Connected Car: Q&A with Danny Shapiro

Especially since it was being run by, not one, but two Tegra visual computing modules (VCM) from NVIDIA, a company best known for powering hardcore computer gaming rigs. Brendan McNally talks to Danny Shapiro, senior director of automotive for NVIDIA to find out more.

TU:  Imagine everyone's surprise to discover the Tesla Model S' infotainment and cluster are being run by gaming processors. What's that all about? How did a gaming chip maker suddenly get into the connected-car space?

DANNY SHAPIRO:  Actually, we've been building in vehicle computing systems for automakers for ten years now. We leverage our graphics processing units (GPUs) from the PC and tablet worlds, but make them automotive grade. They need to be able to stand up to the harsh operating environments and reliability requirements inside a car.

We started working with Tesla about five years ago. They were obviously doing a lot of prototyping early on and the production schedule on that vehicle coincided very nicely with the introduction of our Tegra processors that were coming to market in tablets. An incredible industry milestone was that only a month before Tesla started shipping their cars in the summer of 2012, Google had started shipping their Nexus 7 tablets that also had Tegra processor in it. 

The Tesla Model S has the Tegra system-on-chip in it. Actually there are two.  One drives the all digital instrument cluster, the other the 17" touchscreen navigation and infotainment system.  It was really the first time that consumer electronics technology was in lockstep with the automotive industry, which usually lags behind by several years.

So what we've done is developed a modular solution for our automotive customers.  Tesla has been using it. Audi has adopted it throughout their the Audi line and it's now moving through Volkswagen group vehicles. This Visual Computing Module isa full computer for the car that's programmable. Tesla has done a great job sending out software updates for the vehicle, that basically reprogram the system to add new capabilities, new features, and update the interface.

TU: One of the problems plaguing the whole connected car space is that OEMs and telematics makers are on such incredibly different product development cycles. Being tied to an automaker's three-to-five year development cycle generally means that most embedded telematics systems are verging on obsolescence by the time the car hits the showroom. Are you suggesting you can change this?

DANNY SHAPIRO:  There are a couple things at play here.  One, you are often dealing with car companies who are trying to pinch pennies and so they buy the cheapest computing power, they can get their hands on. Like you said, there's a long development cycle and it can be years before the actual systems end up in the car. So when it finally does, it is woefully underpowered and out of date.  

What we've been doing is working with automakers to time the introduction of these cars so they'll have the state-of-the-art technology that's going to be available at that time. We give them a path to get started on their development work today, so if the car's not shipping for a couple of years, they don't have to lock in today's processors.. Our automotive partners are working on their software while we're devising new hardware. That's one aspect.

Another is, we've created this modular design that is architected, much like your PC where you would update it by popping in a new CPU, new graphics card and memory. Now we are doing that in the vehicle. The VCM gives OEMs the ability to shorten the design cycle. The automaker can iterate each year if they want at the factory as new cars are rolling off the line. That way you don't have to redesign the whole system, you just upgrade the module. 

The potential exists for the automakers to evolve their hardware so that it can be updated at the dealership. In addition, by making the systems programmable, by treating the system like a computer as opposed to just a piece of electronics, gives OEMs the ability to add the features to the capabilities over time and so the car actually gets better the longer you own the vehicle.

TU: But to make that happen, you're going to have to go head-to-head with OEM bean counters. NVIDIA is not what anyone considers a low-cost chipmaker. How are you going to make them come around?

DANNY SHAPIRO:  You're right, we're not a low-cost chip supplier. We're a computer company and a true development partner to our automotive customers. When we deliver a processor, that is not the end of a transaction, rather it is the beginning of a relationship where NVIDIA engineers work side by side with our Tier 1 and OEM partners to develop software. In the auto industry there's an importance in delivering excellent experiences, regardless. The fact is our technology has proven to be driving the best systems in the industry: infotainment and cockpit solutions in Audi and BMW; but it also goes into Volkswagen, SEAT, and Skoda in Europe. We just announced at the Paris Auto Show that our newest customer is Honda.  The Android entertainment system in a Honda is powered by NVIDIA.  So more and more mainstream automakers are recognizing the value of having the best technology inside their vehicles.

TU:  What advantages does a visual computing company have trying to compete in the connected-car space?

DANNY SHAPIRO:  NVIDIA is bringing in a whole new level of computing into the car that's never been there before. Modern cars could have eighty or more ECUs, covering all different functions. in The supercomputing architecture we are delivering with our Tegra SoC can exceed all of those other processors combined. As we add more cameras and sensors to the car, we need to deliver massive amounts of computing power to interpret that data. The car is becoming a rolling supercomputers. 

A lot of people are surprised we're leading Automotive innovation, but that has become clear as for the last four years at the Consumer Electronics Show we are showcasing the most advanced cars  in our booth. The reality is that we have more than six and a half million cars on the road today that use NVIDIA processors for infotainment system, rear-seat entertainment system or digital instrument clusters.. We've developed an all-digital instrument cluster that's configurable by the driver. Moving forward, we're expecting to ship another 25 million more processors to automakers during the next five years.

While much of that will be powering the graphical displays inside the car, an exciting new area of business for us is in driver assistance. Having NVIDIA processors inside a vehicle that are able to interpret camera and sensor data, will make cars of the future much safer.

For the latest telematics trends, check out Consumer Telematics Show 2015 on January 5 in Las Vegas, Connected Fleets Europe 2015 on March 10-11 in Amsterdam, Telematics India and South Asia 2015 on April 13-14 in India, Insurance Telematics Europe 2015 on April 14-15 in London, Insurance Telematics Canada 2015 on April 23-24 in Toronto, Telematics Berlin 2015 on May 11-12 in Berlin, and TU-Automotive @ Detroit 2015 on June 3-4 in Novi, MI.


Leave a comment

Your email address will not be published. Required fields are marked *