SVG
Commentary
Forbes

On The 'Edge' With Driverless Cars

Cameras are visible on the exterior of an Nvidia self-driving car inside the Nvidia booth during CES 2019 at the Las Vegas Convention Center on January 8, 2019 in Las Vegas, Nevada. (Justin Sullivan/Getty Images)
Caption
Cameras are visible on the exterior of an Nvidia self-driving car inside the Nvidia booth during CES 2019 at the Las Vegas Convention Center on January 8, 2019 in Las Vegas, Nevada. (Justin Sullivan/Getty Images)

p(firstLetter). How soon will it be before you pull up at a stop light in your car, and the car next to you has no one behind the wheel? Maybe sooner than you think.

One of the most interesting conversations I’ve had in recent months is with Bryan Mistele, founder and CEO of INRIX, the Seattle-based company that pioneered managing traffic by analyzing data from vehicles as well as road sensors. INRIX is now one of the leading providers of real-time parking and traffic information to your automobile—and they’ll be doing the same with driverless cars, or as the industry prefers to dub them, autonomous vehicles (AVs).

Mistele reminded me that AV's are already on the road. Las Vegas has an autonomous shuttle that opened in 2017, and Denver launched one near Denver International Airport in January 2019.

Talking to an expert like Mistele makes you realize there’s nothing magic about AVs. It’s all about sensors, data, computing power, and connectivity—and ultimately the security of that data and those connections, which can either take us all to the next level of a brave new technological world, or plunge us into chaos.

For example, the experts talk about the transition to AVs is in terms of levels. Level 1 includes built-in “autonomous” capabilities like adaptive cruise control; Level 2 is where two or more autonomous functions (for example, adaptive cruise control plus auto emergency braking) come together in a single vehicle.

At Level 3 the car still needs a driver to keep eyes on the road, but the car handles most tasks by itself, for example highway driving, and even can do the principal driving to get you from home to work and back. That’s where most of us will be in the next five years.

At Level 4 the car won’t need a driver except under certain conditions—for example with changes in the weather or road conditions like black ice or a gravel surface which the typical AVs’ optical sensors have trouble detecting. Most self-driving vehicles being tested today fit this description, while companies like Tesla continue to push on to Level 5, the nirvana of the AV engineer. That’s where the car does it all under all conditions, with you being the passenger the way your “baby on board” is now: free to read, sightsee, or sleep without giving the actual driving a second thought.

Think that’s farfetched? BMW and Volvo say they’ll have a model capable of Level 4 or 5 by 2121. That may be corporate hype; but what will make that reality depends on far more than just the car itself.

First of all, it will depend on how well the avalanche of data that flows in from an army of optical sensors on board the car, can be analyzed and understood.

When I say avalanche, I mean it. Today’s “smart car” like the BMW 3series comes with about 50 sensors processing 50 megabytes of data a day. By contrast, the average AV will have 400 to 500 sensors, processing 40 terabytes of data every eight hours of driving. That would translate into more than 15 billion single-spaced typed pages of data.

Sorting that volume of data will require very sophisticated computing power: what’s called “edge computing,” meaning doing the workload inside the car itself instead of relying on a removed data center. This means every AV will need on board the equivalent of a squad of supercomputers. Fortunately, a company called Nvidia has already worked that out. Their PX Pegasus drive control system is able to handle 320 trillion operations a second, which will take your car to Level 5, with a computer the size of a license plate.

But making AVs a reality will also require one more technological feature, which is connectivity, and for that we’ll need 5G. Instead of relying on a connection to the cloud, 5G will allow your AV to process data directly from nearby vehicles and from vehicle infrastructure communications, for example stop lights that send their signals to the car and road signs that pass on information about traffic patterns and speed limits. Las Vegas is the first city to offer fully digitized rules of the road that an autonomous vehicle can directly engage with, and thereby learn which streets are one way, what the speed limit is—and also where parking is permitted that day and where it’s not.

Under the AV big tent, in other words, 5G will be the ringmaster. It’s the technology that will permit the free and instantaneous flow of data that will make riding in an AV a matter of routine, rather than an adventure, and supply the connectivity that an AV will need to coordinate with other vehicles, whether equipped with human drivers or not.

If that connectivity works, then AVs will transform our lives for the better. If it doesn’t, then the AV experience will become a nightmare.

For example, if a malicious actor can tap into the information communication networks that coordinate and direct the flow of AV traffic, he or she can bring a city—even a country—to a halt. This means having future 5G networks we can trust. If people are looking for another reason why having a company like Huawei run 5G is a bad idea, here it is. Given the company’s malicious track record as a cyber thief and tool of the Chinese military and intelligence agencies, handing the future of AVs over to Huawei seems a risk not worth taking.

The first seventy years of the computer revolution were primarily about power: getting machines to perform operations faster, smarter, and more efficiently—and part of the advent of AVs is certainly about that. Today’s era, however, is about security; making sure the data and networks we rely on every day are safe and free from intruders and malicious actors, whether it’s our laptop, our smart phone, or—by 2024—our self-driving car. The old adage that an ounce of prevention is worth a pound of cure couldn’t be truer. When it comes to securing the future of 5G, that ounce of prevention could end up making the phrase, “Leave the driving to us,” part of the American Dream instead of an invitation to gridlock.