Mass production and widespread use of automobiles in the early 20th century radically transformed the industrialized world. But according to Mamatha Chamarthi, CIO of Fiat Chrysler America, the converging megatrends of automotive autonomy, connectivity, electrification, and sharing (ACES) will have an even greater effect on how we live, work, and interact. This post is about one of these megatrends—automation. We’ll explore both how and when your car will be fully capable of operating itself in all driving conditions, without the slightest intervention on your part (what the Society of Automotive Engineers classifies as Level 5: Full Autonomy).
As with so many other industries, we can, in general, credit the ongoing digital revolution as the driving force behind the disruption of the car. Helping to make autonomous vehicles safe, effective, and affordable, though, will require the synthesis of three engines of change, in particular. These are:
- Enormously powerful computer processing
- Inexpensive, versatile data storage options
- Low latency, high-speed connectivity, such as 5G when it reaches its full potential
As these three forces activate each other, the engineering and widespread adoption of fully autonomous vehicles could shift from concept to reality. Chamartha thinks this will happen in the 2030 to 2035 timeframe, and describes the CIO role in bringing it about as “mind-blowingly exciting.”
Sensors & sensibilities
When we think about what’s happening under the hood of a driving car, images come to mind of pistons, gears, camshafts, and other parts and processes working in concert to propel the vehicle forward. But in today’s automobiles, which can have around a hundred million lines of code, the importance of software engineering is starting to overtake that of mechanical engineering. As we move closer to Level 4 or 5 autonomy, this will be even more pronounced. In Charmarthi’s words, “for a car to have complete autonomy, we’re talking 300 million lines of code.”
In part, all those lines of code are necessary to operate the stunning number of sensors that are in modern cars and, to a much greater extent, will be in future cars. Increasingly, cars are becoming mobile, multi-sensor devices. In the coming years, your car’s sensors will communicate with sensors embedded in everything from other cars and the road to the infrastructure of your city or town. One day, this constant, near real-time 5G communication, termed Cv2x (cellular vehicle-to-everything) will produce a staggering amount of data. By some estimates, one car will generate 4TB of data per day—data about your car, the environment through which it travels, and yes, you. Automobile OEMs will increasingly rely on this data to find optimal ways to maintain and even enhance operations by identifying and bringing about new business models.
Chamartha illustrates this point using an easy-to-follow example. A car’s sensors, she posits, could one day collect and analyze data that indicates your brake pads are wearing thin. The car could then alert you to the condition of the pads and automatically order replacements. Finally, the car could ask if you’d like to have the pads installed at the dealership you’ll pass on your way to work or, for an additional fee, in your office parking lot as you go about your day.
In this scenario, everybody wins. The car OEM or parts manufacturer and the dealership get more business while you and everyone else on the road drive more safely as the possibility of an accident due to your improperly maintained brakes diminishes.
Though we’ve been hearing about the promises of artificial intelligence for a long time, where automobiles are concerned, Chamarthi says it’s becoming closer to reality. As mentioned earlier, a car’s many sensors can communicate with other cars, the road, traffic and weather monitoring devices, and so on. This communication produces vast amounts of data which in the future could be moved via 5G to the cloud and, by means of powerful processing, analyzed for patterns. These data patterns won’t merely be historical, they could also be predictive and even prescriptive. The patterns that emerge from the processed data could guide the decisions and resulting actions of the car’s A.I. With the future of 5G, all this communication, as well as analysis, decision-making, and autonomous action, could occur on the fly in what is effectively real time.
What does this mean in the real world? Imagine your car, for example, could quickly apply the brakes when the bus in front of you stops suddenly to avoid hitting a dog (which, fortunately, has a sensor in its collar). It’s only when all the necessary processes converge—data collection, communication, storage, analysis, and action—that safe and effective autonomous cars will become a reality.
It can’t happen fast enough
Some 1.2 million people die in automobile accidents every year, 95 percent of them as a result of human error. A future where 5G, A.I., and, going forward, computing at the edge could potentially augment human capability enough to start helping to save lives on the road is something Chamarthi is especially keen on effecting. Getting the technology right is a start, but even Level 5, fully autonomous cars won’t go anywhere until governmental regulations allow them, people have the confidence to get in them, and more than just billionaires can afford them.
Then, when everything’s finally in place, the fully autonomous car will arrive. Ready to go for a ride?