I think this gets at the heart of the “driving towards driverless” question… Vehicles that fall into SAE levels 1-4 are available today, but the leap to true fully automated driving is non-trivial. See Wired article here for SAE levels definition. Every idyllic image of driverless vehicles, including people sleeping or working, requires full automation. Many of the stated benefits of driverless vehicles also require full automation. Waymo is likely the company closest to this reality, but many automakers are stating that they’re working towards full autonomy in the next few years. Should we believe them?
So what’s keeping us from full autonomy? Let’s ignore the lack of regulations and societal acceptance. The number of unpredictable and/or complex situations are, literally, endless. Fallen trees post-thunder storms, an overpass collapsing, construction work zones, traffic detours, and the list goes on! Our society seems to gauge progress towards Level 5 autonomy based on number of driverless test miles “driven;” however, how many miles will enable driverless vehicles to predict all of these extreme situations?
I believe that the more “connectivity” we have, the sooner full autonomy can come. “Connectivity” refers to all aspects of vehicle to vehicle (V2V), vehicle to infrastructure (V2I), and vehicle to everything else (V2X). Connectivity is going to increase the amount of information available to share with driverless vehicles and minimize the risk of those unpredictable situations. When will more V2X applications be available? This will require a significant investment in standards, infrastructure, privacy and data sharing policies, and cybersecurity protections.
Bottom line – I am not expecting to see fully automated (level 5) vehicles for a while, but what do you think?
Pingback: Will We Ever See Level 5 (Fully Automated) Driving? | Car News, Reviews, & Pricing for New & Used Cars.
I agree with you Lauren. And, as a member of the work zone safety community, I hope they take their time. Our work zones are by their nature unpredictable. Today, many level 5 test vehicles simply stop when they encounter something unexpected. Do that in one of our work zones and it will cause a chain reaction of rear-end crashes. Better to test their technologies on closed tracks first.
I would also add we are doing our part by creating new work zone technologies that will update the digital map infrastructure with work zone locations and changes in geometry in real-time. That will add to that connectivity you described!
Great article Lauren and I agree. I think the challenge as you note is how do AVs get past what I call “The Other 1 Percent”: those infrequent, but most difficult of driving conditions that will arise and are generally thought to require flexible, advanced level judgment. For example, a riot or other rapidly evolving situation. Even if they only occur once 300,000 miles, the car needs to be able to handle them.
I think an interesting discussion is: What if we are never able to achieve truly Level 5 vehicles because of these unique situations? What implications will that have for the other predictions and planning about AVs?
I think that rather than send lower level technology into the public arena like TESLA has done with inevitable mishaps, Level 5 technology should be experimented with in isolated and cordoned off conditions. Lets say, several square miles of Detroit, or even a place outside the United States where liability is less of an issue, and pay a lot of good people to figure it out. It wouldn’t be cheap, but I think in the long term a lot cheaper and quicker than trialing experimental technology in incremental steps in the real world – which with occasional mishaps can only set the whole thing backwards. Then once it is up and running, can be showcased around the world as a leading example of how it can be done. Not incomparable, I think, to the Saturn program that got the United States to the moon in just a decade – it would take that level of financial and political commitment. A risky business it must be said, but the rewards could be very great.