Autonomous Vehicles: Six Slow Months

I last published on this topic June 28^th, wrapping up the previous 12 months, and July 11^th, with projections on the shares of vehicles reaching certain technological points in the USA and elsewhere and the number of trucking and cabdriving jobs in relation to today’s.  I have established that forecast as a regular annual feature, but didn’t want to get caught up in the often-daily updates in the field I had seen for the previous year.  So what has happened since then?

One noteworthy thing is the lack of news itself.  Companies have been much quieter about their progress, with the largest stories concerning business instead of technical moves.  The promises have softened, especially, as we will see, those from the formerly loudest participant.

One company reached a real milestone in the summer, as “Lyft’s self-driving vehicles have performed 5,000 passenger rides in Las Vegas” (Cohen Coberly, TechSpot, August 22^nd).  Although all were in the limited and almost linear area of the Strip, it’s still an impressive accomplishment, especially with no accidents and a stunning 4.96 out of 5 passenger feedback record with comments calling the trips “amazing” as well as safe. 

On the regulatory front, our federal government showed again that it can be helpful and set guidelines without unreasonable or heavy-handed interference.  In “US Department of Transportation updates autonomous car rules,” from Engadget on October 4^th, Natalie Behring showed us changes to voluntary, not mandatory, principles, including allowing automata to legally constitute “drivers” and announcing an intention to drop requirements for devices, such as pedals and steering wheels, such systems don’t need.  That approach has drawn disagreement from the private, consumer-advocating Center of Auto Safety, but is certainly the long-run winner.

I thought we had more insights into the recent slowdown in “Through All the Hype, Self-Driving Cars Remain Elusive” (Norman Mayersohn, The New York Times, November 27^th), but this article, after noting the likes of “self-driving vehicles, despite being the subject of breathless media reports and in automakers’ strategies, remain years from being available to private owners” (which we know, as the first wave of them will be put in fleets instead), and quoting a Stanford transportation lecturer as saying that “few start-ups actually understood the commitment required to create a complete vehicle” (but none in the past year have tried), listed companies not only specializing in different technical areas but taking varying business approaches, from concentrating on easy environmental conditions to planning to deal with all of them, from partnering with one automaker to offering products to all, and starting with taxi service, intracompany runs, or other fixed-route shuttles.  Mayersohn revealed ten of these concerns, some familiar from the literature but several not, which showed how much forceful, widely-varied work is still ongoing.

In the December 5^th Arizona Republic, Ryan Randazzo told us about his progress-assessing project, “We followed Waymo’s self-driving cars around Arizona for 170 miles:  Here’s what we saw.”  The latter, though clouded by his uncertainty on whether the vehicles, all of which had safety drivers, were actually in driverless mode, included extreme caution while approaching a major accident site, perhaps excessively slow turning in some intersections, sluggish lane changing causing some missed turns, and what seemed like unusually conservative driving in general.  Randazzo’s findings cast a positive light on Waymo, which has already shown itself to be one of the soberest and most measured autonomous-technology providers – it seems appropriate to err on the side of caution, and problems such as not getting into the proper lane in time will clearly be attacked and solved.  Remembering that all know that such vehicles are not at all being touted as finished or even commonly available should remove any concern about this piece’s discoveries.

That company recently crossed another line, as “Ex-Google driverless car firm Waymo begins charging for self-driving car rides in Arizona” (USA Today, also December 5^th).  It is not a general offering, but only for “pre-approved passengers in the Chandler, Arizona area.”  Meanwhile, as I alluded to before, “Uber plans smaller, more cautious self-driving car launch” (Heather Somerville, Reuters, again December 5^th).  This is the first sentence I’ve seen with both “Uber” and “cautious,” but seems to be what came off their drawing board after their real but grossly overemphasized fatal March accident.  Their tests will resume at speeds under 25 miles per hour in dry daytimes with not one but two “employees” in front seats, with no plans for passengers and “no firm start date.”  That may be so unadventurous as to make Uber uncompetitive.  However, Somerville’s reporting of this announcement was more positive than what Rachel England wrote in Engadget’s December 6^th “Uber puts self-driving cars back on the road in scaled-down test,” including the statement that “current employees have anonymously claimed that Uber is taking shortcuts to hit internal milestones.”  If that were to be documented and to have demonstrably bad consequences, it could finish Uber off, not only in the autonomous-vehicle realm but eventually as a business.  It is now more important than ever that we differentiate between what Uber is doing and the progress and safety of others. 

Four conclusions stem from the events above.  First, the limited rollouts and smaller set of immediate future implementations mean that we are behind schedule, maybe six months back of what my July projections anticipated.  Second, the Waymo model of operating in smallish, well-mapped and defined areas which can expand with time may become established as the way driverless technology reaches the general public.  Third, the wide variety of companies and methods is good for the long run, as some will be successful and most will not, but may mean further delays in the next year or so.  But fourth, if people working in this field maintain or reimplement their 2017 levels of intensity, there can be no doubt that driverless vehicles will, indeed, become the norm.