IT innovation: Paving the road to self-driving cars

Typography
Features

Jack Pouchet, vice president, Market Development, Vertiv and Jodie Bare, deputy program manager, Smart Columbus, explore why the change to transport as we know it may well be imminent. 

After just three turns in any given chess match, there are 121 million possible outcomes. Think about that. Two players, 32 pieces, on a board of 64 squares. And just three turns into the game, there are roughly as many potential outcomes as there are people in Japan. 

Now think about your drive to work. The average person’s commute is about 25 minutes. How many data points are processed in those 25 minutes? We’re constantly calculating and adjusting to countless variables – speed, road conditions, traffic signs and signals, pedestrians, other cars on the road, even animals that dart in and out of traffic. 

Every decision we make – those we consider and the others that take place virtually subconsciously – adds countless possible outcomes to our daily drive. That we manage this every day, mostly without incident, is a marvel of the human brain that we take for granted.

Now consider this: More than two decades after a computer named Deep Blue defeated world chess champion Gary Kasparov in a game of chess, we are on the verge of a similar breakthrough with our cars. Self-driving vehicle pilots are taking place in 19 cities, with infrastructure and industry insight expanding daily to enable future deployments. 

Columbus, Ohio, is one of those cities. Smart Columbus is using a $50 million grant from the US Department of Transportation (USDOT) and $522 million more in private-sector investments to explore and evaluate advanced transportation systems, including self-driving vehicles. 

The pilot sponsored by Smart Columbus and Drive Ohio involves a self-driving shuttlethat transports passengers to and from popular downtown destinations. It launched in December and is an educational milestone on the path toward future deployments in Columbus neighbourhoods and beyond. 

The infrastructure challenges associated with an autonomous future are real, however. All of those variables and decisions we make during that 25-minute drive to work require constant data transmission and analysis and real-time adjustments. That requires sophisticated computing not just in the vehicles, but along the roadside – every roadside, everywhere. 

Autonomous vehicles will generate almost two terabytes of datajust on that drive to work. That means on your daily commute, your average driverless car will generate data equivalent to about 2,000 movies. That’s every car, every hour. 

The IT systems needed to make that a reality must be incredibly powerful, small enough to mount on light poles or stored in ubiquitous, nondescript roadside enclosures, capable of low-latency transmission, and uncompromisingly reliable. That last part requires power protection, thermal management and, importantly, remote monitoring, management and service capabilities. These sites will make up a micro-edge, capable of instantaneous computing and networked not just with each other but with data centres and cloud resources around the world. 

Smart Columbus is working with a number of partners to find solutions to those infrastructure challenges and other speed bumps on the way to a driverless future. One of the most powerful tools in that effort is the Smart Columbus Operating System, an open source data platform that provides open access to the city’s latest mobility data. 

The operating system aggregates data from a wide variety of sources, allowing Smart Columbus to glean insights from across the community and from outside resources through USDOT. It is revealing important data about driver behaviour, traffic and parking patterns, and technology performance, all of which will contribute to a more effective strategy for implementing self-driving vehicles. 

For its part, Vertiv is engaged not just with Smart Columbus, but across the industry to develop intelligent, reliable infrastructure solutions that support 5G and edge networks. Those efforts will enable self-driving cars, Smart Cities, virtual and augmented reality platforms, and widespread applications of the Internet of Things. 

As much as the car is the natural focus of so much discussion on this topic, the reality is this is a data centre and IT challenge that carries with it all the individual issues driving data centre innovation every day. Those include: how to effectively store, analyse and move mountains of data, how to reduce latency and move that data quickly, and how to ensure those systems that are becoming absolutely mission critical remain online at all times and in all conditions.

Bottom line: We’re moving closer to an autonomous reality. Data centres and advanced networks will make it possible, and our Deep Blue moment is closer than you think.