Cars Unlikely to Become more Computer than Transport

The automobile industry wouldn’t be as sprawling and as massive as it is today without effective marketing hype.

Recently, advances in ADAS technology have some individuals in, and adjacent to, the industry saying that vehicle “brains” will approach, or even match, the status of supercomputers. Let’s take a quick glance at this myth before delving into the more interesting reality.

S.A. Applin, an anthropologist and research fellow at hrafARC, who concentrates on the impact of technology on society, said bluntly that this is impossible. “Supercomputers are exceptionally advanced apparatus. They can’t fit into a car physically and they can currently cost over $500M,” according to her.

That being said, the power and scope of in-vehicle systems is indisputably, and endlessly, growing. No, we won’t ever be riding atop monster computers dedicated to carting us around autonomously but where might this ever-accelerating pace of innovation get us to?

Some speculate that in-car systems will become too powerful to leave idle and are planning accordingly. Canadian next-generation vehicle maker Daymak says it is packing cryptocurrency mining capability into its in-car systems, which will kick in with the 2023 Spiritus, a three-wheeled EV currently in pre-order. According to Daymak, its Nebula software platform will allow owners to not only mine cryptocurrencies but store them on the company’s proprietary crypto wallet. So, in other words, the Spiritus could theoretically mine and stash happily away while parked and awaiting owner in the driveway, the garage or on the street.

Maxwell Zhou, CEO of robo-truck and taxi developer DeepRoute.ai, is one executive knee-deep in ADAS who’s skeptical about this outside-the-car computing potential. He pointed out that “at this moment, it would be ideal if we could crowdsource the spare computing power and apply it to other functions but for, highly automated vehicles, the computing power is positively correlated with the cost so, usually, there won’t be much spare computing power left”.

As we advance up the levels of automation, the need for greater computing power grows exponentially. “Cameras and sensors rely on high computing capacity to process large amounts of data on the road in real time. The planning and control algorithms also require high computing to handle complicated road scenarios to ensure safety and efficiency. In addition, domain controllers and OTA upgrades also need support from computing power,” Zhou said, ticking off only a few of many examples.

Another hindrance to expanding an in-car system outside the confines of the vehicle is the consumer’s expectation and his or her choices and preferences. “Surely people will be driving and not computing,” Applin said. “For passengers, the issue becomes why the car as a tethered device would matter, when people can easily bring their mobiles and laptops wherever they go?”

So, perhaps rather than pushing hard to approach or imitate the speed and effectiveness of true supercomputers, ADAS and (eventually) AV solutions developers would be wise to build out their driving functionalities as much as possible. Although Nvidia, an in-car systems specialist, is one company that likes using the “supercomputer”, its executives don’t envision a world of parked autos mining cryptocurrencies or modeling weather patterns.

The company’s vice-president of automotive Danny Shapiro, said: “Building AVs is one of the most complex compute challenges of our time.” Since a self-driving solution is so mission-critical, “these vehicles will need to be richly programmable and support software updates OTA, delivering new capabilities and features for greater safety, functionality, convenience and overall value to the consumer over the life of the car,” he added.

They’ll also need to continue devoting a great deal of computing power, no matter how much they scale, to the functionality most prized by consumers – safety, of course. A 2019 poll conducted jointly by news agency Reuters with market researcher and consultancy Ipsos found that two-thirds of respondents believed AVs should be held to stricter safety standards than those of the human-driven variety.

For their part, some authorities are listening and acting on this wish. This past March, the US National Highway Traffic and Safety Administration published its “final rule” on AVs, an exhaustive, and exhausting, 155-page document with a dizzying number of thresholds for everything from airbag operation to parking brake positioning. These are only the requirements set out by one country’s road safety authority; it’s very possible that consumers around the world will demand even higher standards. So, ADAS and AV solutions developers should devote plenty of that ever-advancing in-car computing power to functionalities that provide at least as many safety functionalities as the public clamors for.

HRAF’s Applin, for one, cautioned that developers should never lose sight of the end user; at least some research capacity should be devoted to their wants and needs. “If manufacturers want to know what consumers want, they should hire anthropologists to conduct qualitative research studies to uncover underlying consumer needs,” she said. “Surveys and speculation don’t get as far because they presume to know potential outcomes versus discovering them.”