It's time to democratize safety. Sure, the tech industry preaches about safety and highlights the failings of humans as drivers. But let's take a closer look at what's really going on here...
I believe it’s time to democratize safety. The tech industry preaches about safety, highlights the failings of humans as drivers and boasts of autonomous driving to save lives. But let’s take a closer look at what is really going on here.
It is so easy to write off humans as bad drivers. Aren’t 94% of accidents the result of human error? Don’t human drivers kill more than 1.3 million people globally on our roads each year? Isn’t the obvious solution to get human drivers out of the loop and replace them with machine drivers? Are you serious?
We’ve got the safety arguments back-to-front. Take a step back from the hoopla, from the marketing BS and from the unfounded promises of autonomous driving and just ask yourself this: Where is the mandate for proven safety technology for human-driven, mass-market vehicles? If that is the low hanging fruit, why aren’t we starting there?
Let’s begin with the National Highway Traffic Safety Administration (NHTSA), a US agency of the Department of Transportation (DoT), which last December released its advance notice of proposed rulemaking (ANPRM) for autonomous vehicles.
Autonomous vehicles make for really exciting reporting, but experimental technology belongs in Silicon Valley, not Washington. Safety policy begins with what is validated, what is verified, what is affordable and what is realistic. If you’re going to San Francisco, be sure to wear a flower in your hair. If you’re going to write transportation safety policy, start with what works. Safety advocacy sounds more like Robert Sumwalt, chairman of the NTSB, than Elon Musk, Anthony Levandowski or Travis Kalanick.
Where is NHTSA’s mandate for basic life-saving automotive technology? Yes, humans get distracted and drowsy. Our minds wander on to other things. Driving is mostly monotonous, but life-threatening events can, and do, occur on public roadways without warning. Motorists, cyclists, scooter riders, pedestrians and other road users come together to share the same limited space, sometimes chaotically.
Humans have great perception and can navigate complex situations far better than any AI or deep learning robo-driver. But we also have a neuromuscular lag of about half a second, meaning from the precise instant you decide to hit the brakes to the moment your foot actually moves is equivalent to a distance traveled of about fifty feet at highway speed.
Machines react faster than humans in an emergency. So why isn’t installation of automatic emergency braking (AEB) covered by a NHTSA mandate? https://www.iihs.org/news/detail/10-automakers-fulfill-automatic-emergency-braking-pledge-ahead-of-schedule If AEB can save lives, then the most sophisticated variant of AEB (called AEB-VRU standing for Vulnerable Road User) can save more lives. So why isn’t the installation of AEB-VRU covered by a NHTSA mandate?
If unintended lane changes cost lives, where is the mandate for automated lane keeping systems? If distraction and drowsiness lead to fatal accidents, where is the mandate for vision-based driver monitoring systems? Or blind-spot monitoring?
Why does investigating the possibilities of test-level autonomous driving technology have a higher priority at NHTSA than mandating the installation of simpler, cost-effective, technology or updating the NCAP (new car assessment program) safety ratings, either of which would immediately reduce road deaths? Safety on public roadways either matters or it doesn’t, and NHTSA’s actions look more like an agency in thrall to the tech industry than an agency taking its remit of traffic safety seriously. Something, somewhere, has gone wrong.
Faith no more
It is 2021 and my faith in nearly all of the autonomous driving companies has gone. Writing on Twitter recently Kyle Vogt, co-founder, president and CTO of Cruise boasted:
Cars are driven by humans. Humans improved safety by ~2x in ~50x years. We suck. Robots are likely to improve this by 100x or more in 3-5 years. That is a big deal. Car accidents are the #1 killer of teenagers in the US, a very uncomfortable fact.
I’m calling that what it is: Marketing BS. We don’t need partisan, unfounded, hype from Vogt to make public roadways safer. We need a mandate to install effective accident prevention technology on mass-market human driven vehicles as soon as is practically possible.
This report in Techcrunch summarized the recent $2 billion investment in Cruise by GM, Honda and other institutional investors including Microsoft.
GM’s CEO Mary Barra needs to have a blunt conversation with Vogt and tell him to get the flowers out of his hair and stop making crass and boastful statements talking up the promise of Cruise. Flash words won’t save lives.
Automakers pouring billions into autonomous driving technology all the while foot-dragging on the installation of basic safety kit on their standard human driven vehicles have reached the end of the road. They can hide their hypocrisy behind their hoopla no longer.
Also writing recently on Twitter, Waymo CEO John Krafcik commented:
Roads would be safer if we could agree on a simple taxonomy & use it consistenly (sic): If a licensed human driver is required, it’s driver-assist technology. If no licensed human driver is required, it’s full autonomy.
Seriously, John? This even makes your “to-do” list? I would argue that roads would be safer if every CEO of an autonomous driving company publicly demanded that NHTSA mandate AEB, lane-keeping systems and vision-based DMS in all new road-going vehicles with four or more wheels. There is even a case that vision-based DMS should be retrofitted to every vehicle weighing 10,000-pounds or more, to permanently monitor for distraction, drowsiness and impairment in drivers on public roads.
The safety arguments in the U.S. have ended up back-to-front. The priority needs to be to make human drivers into safer drivers. Let’s do that first and then see if the market wants or needs autonomous driving technology. This is the route already being followed in Europe and there are early signs the Biden Administration may now be heading down a similar path.
I can make a great suggestion where to begin. Pete Buttigieg, the new secretary of transportation under the Biden administration, could start by filling all of the vacant senior positions at NHTSA, as shown by the January 2021 organization chart, as fast as possible.
The democratization of safety
This week, Qualcomm is set to make some major announcements related to automotive technology. Reviewing division head Nakul Duggal’s presentation to the Automobil-Elektronik congress from last year, we can see that Qualcomm looks to have integrated driver monitoring as standard into its cockpit application processor and longitudinal speed assist (ACC) and lateral lane support (LKA) into its ADAS application processor.
We will know more details about Qualcomm’s plans in a couple of days, but we appear to be at the dawn of what others have termed the democratization of safety, where the installation of critical life-saving technology rapidly becomes ubiquitous across the automotive industry and not limited just to luxury and high-end models.
It is the coming together at the system level of driver monitoring and ADAS technologies which so interests me. Previously these were two discrete technologies that had to be integrated at either the tier 1 or automaker level, and neither Nvidia nor Mobileye viewed the adoption of state-of-the-art driver monitoring as a priority. In automotive safety, as in smartphones, Qualcomm looks to be starting out with a system-wide approach to solving all of its customer’s problems at once.
In the next week we could have heard announcements from several major automakers to fit ADAS and driver monitoring as standard on all their series-production models, potentially covering tens of millions of vehicles. That will throw down the gauntlet to the remainder of the industry and as investors in Cruise, GM and Honda need to be in the vanguard of the trend.
For all the talk of autonomous driving and saving lives, the greatest safety advances might be about to come using technology from a company that none of us forecast would drive this revolution: Qualcomm.