Self Driving Reality: It's Here, In Snow, In Unexpected Conditions, Now

Waymo Celebrates 10 Million Miles of Self-Driving. Can you drive better than the video below?

Waymo says "We’re building the world’s most experienced driver, with over 10 million miles self-driven on public roads and almost 7 billion in simulation. Discover what we've seen and learned along the way from literal curveballs in our path, to navigating low-visibility dust storms, and more unexpected scenarios."

Self-Driving Car Timeline

Tech Emergence provides Predictions from the Top 11 Global Automakers

  • Ford Motor CEO Mark Fields told CNBC that Ford plans to have a “Level 4 vehicle in 2021, no gas pedal, no steering wheel, and the passenger will never need to take control of the vehicle in a predefined area.”
  • According to Reuters, GM is rumored to have plans to deploy thousands of self-driving electric cars next year with its ride-sharing affiliate Lyft Inc.
  • At the end of last year Honda announced it was in discussions with Waymo, an independent company of Alphabet Inc., to include Waymo self-driving technology in their vehicles. The long stated goal of Honda is to have cars that can at least drive themselves on highways by 2020. That is when Tokyo will host the Summer Olympics, and Japan hopes to make it a showcase of their technological prowess.
  • At the beginning of the year Daimler announced a deal with Uber to introduce their self-driving cars on Uber’s ride-share platform in the coming years. Like several other car makers Daimler view mobility as a service as a logical place to first use self-driving cars.

Those are just a few paragraphs. All the car makers seem committed.

Adoption Rate

Most of the car makers will will all be ready by 2020 or 2021. Some say 2022.

But capability is one thing and mass adoption by the public is another. It may take a few more years for cars.

The article concluded "Even with a heavy degree of skepticism it seems likely that if you live in a major city you will be able to hail some form of automatic car ride in less than a decade."

About 2-3 years is more like it, in most major cities.

“Freightliner Inspiration Truck”

A quick check of my calendar show the date to be November 20, 2018.

Self-driving trucks are legal on Nevada highways with permits. Arizona and Florida welcome self-driving trucks without any special permits or reporting.

Long Road to Self-Driving Trucks

Medium comments on the Long Road to Self-Driving Trucks

The author, Mark Harris, points out some remaining issues and he is also a bit skeptical about earky adoption, but the cost savings alone insures it will happen sooner rather than later.

> Unlike the high-profile urban and suburban tests of autonomous passenger cars conducted by Waymo, Uber, and GM Cruise, driverless truck developers have been quietly carrying out their experiments on highways across the Sunbelt. According to figures submitted to the DOT, automated trucks racked up well over 200,000 miles on public roads last year.

> A truly driverless truck would change the face of trucking. In a cutthroat industry operating on razor-thin margins, drivers’ wages and benefits are the largest cost, accounting for nearly half of a carrier’s cost per mile. The first company that can eliminate drivers could undercut its rivals and dominate the market.

> Automated trucks might also reduce the nearly 4,000 fatalities from large truck crashes that occur each year in the United States, some of which are known to result from driver fatigue. There is even a potential environmental benefit, as smart driving systems reduce fuel use and smooth the transition to alternative fuels and electric vehicles.

Timeline on Schedule

The 10-year timeline I posted in 2012 seems to be on schedule.

Mass adoption will first be with commercial trucks, then taxis, then the public.

I expect national regulations to be in pace by 2020. If so, mass adoption by highway trucking will commence by 2022.

The cost savings, competition, and shortage of truck drivers ensures that outcome, like it or not.

Mike "Mish" Shedlock

Comments (48)
No. 1-24
Jojo
Jojo

I wouldn't want to be a steering wheel manufacturer!

Detroit Dan
Detroit Dan

Well, there's also this:

and

"We're not even remotely close to being able to be truly autonomous in diverse conditions," said Austin Russell, CEO of lidar manufacturer Luminar. Motorists rely on human cues that technology cannot detect: the gestures of a traffic cop, or eye contact with a pedestrian or another driver that can help us predict their behavior." (from https://theweek.com/articles/807864/when-selfdriving-cars-take-over )

No, self-driving cars are not here. It's still mostly hype. As Steve Wozniak says, "I'm sick of the lies."

Please see

ReadyKilowatt
ReadyKilowatt

CDOT is adding more truck parking. This is a reaction to new technology keeping drivers honest about their logs and rest periods. But I have to think a lot of this has to do with poor weather events keeping self-driving vehicles stopped. Vail Pass closes pretty much anytime there's snow which even in a drought year is a weekly event. At a minimum the "chain law," which requires commercial vehicles to carry and often use snow chains on the passes, won't work unless there's a driver. Sure there's automatic chains for smaller vehicles, but big trucks don't have that option. My guess is that as fully autonomous vehicles become common they'll just stop when the "chain law" is in effect.

"Truck stops could also bring prostitution and human trafficking, Gamba said, though he admitted the association was purely anecdotal." - Not if there's no one driving the trucks.

Vitos
Vitos

I would be willing to pay a lot of money for self driving capability in a new car, even if all it could do is cruise on the freeway while I napped. This seems doable in the near future. I'll be happy to take the wheel in cities, snow, etc.

douglascarey
douglascarey

'About 2-3 years is more like it, in most major cities. " Mish will be as wrong about this as he has been about GDP growth predictions. We are at least a decade away, probably 20 years. Why? Because government roads are involved. How long to build that bullet train in CA? 15 more years at least they say. And you think self-driving cars will be here in two or three years? What a nutty prediction. At least Mish is on the correct side with tulips, I mean bitcoin.

KansasDog
KansasDog

How about a combo manual/computer car. You drive like an idiot, tailgate, etc you lose your manual status for 30 days. If you like to speed a lot and are always in a hurry the computer option will go 5mph slower than the speed limit. LOL. Or parental controls. Just saying this out of frustration over the decline of peoples driving habits. Thanks to tech everyone is in fast mode I think its translating to driving. I live on gravel, I have cars fly by at 60mph several times daily. On the local hiways I am consistently being passed in low vision areas including going up hills if im not driving at least 10mph over. Would be different on an interstate but out here you are easy pickens for a ticket. Pisses me off.

Realist
Realist

Self-driving vehicles ARE coming. The only real debate is how soon and how fast. Mish is an optimist and expects them sooner than most. While I wish the technology was going to be ready as quickly as Mish suggests, because I am a big fan, I am less optimistic about how soon driverless will have a big impact. I believe that we are roughly 5 years into the 10 year introductory phase, where we go from 0% to perhaps 0.5% of all vehicles being autonomous. The second decade will take us from 0.5% to perhaps 40% and will be a huge societal change.

Carl_R
Carl_R

In a few years some will say "See, they are available, I'm right", while others will say "There aren't that many on the road yet, so I'm right". Everyone will be right.

Kinuachdrach
Kinuachdrach

The economic opportunity is in self-driving big rigs for inter-city long-distance driving, picking up drivers at depots on the outskirts of town for in-town delivery. That assumes the manufacturers will be able to develop ways to defeat future hackers who will inevitably seek ways to hijack the driverless trucks out on the highway.

Self-driving automobiles? Maybe in a few small niches, and probably quite slowly. One problem is expense -- additional equipment required for self-driving will not be cheap. Add those costs to the already high price of new cars, and unsupportable prices will really cut demand, especially when interest rates return to normal levels. A second problem is maintenance for all those high-tech sensors -- it will take time to learn how long they last in the real world, and how expensive they are to diagnose & replace in actual service.

One other issue is the amazingly improved service life of modern automobiles. It takes a long time to turn over the vehicle parc -- at least two decades. Unsustainable government borrowings, unrepayable national debts, and welched government commitments on pensions & healthcare will have upset the apple cart before those two decades are out.

Schaap60
Schaap60

I've been following this line of posts since Mish has addressed the topic. I've been reading about driverless cars even longer. The drop in people saying it will never happen is noticeable; the argument now seems to be about timeline. Ride hailing in major cities is very different from mass adoption, and even that depends of the definition of ride hailing, i.e., does it require being picked up at home, or does it merely include cars operating on designated routes that can pick you up anywhere along the route? The latter seems more plausible than the former in 2-3 years in major cities, and fits within the definition of the article. That's why I think Carl R is correct.

I agree costs and maintenance are an issue, which is why driverless may take much longer for individuals to adopt for personal use. However, the fact cars are currently parked 95% of the time is also very expensive. I think that's why ride hailing, where the cars can stay in service for longer periods of time in major cities, is much more likely to be adopted sooner so the costs of operating these vehicles can be spread out among more users.

There is also a growing demand for this type of transportation option as society ages and more people can no longer drive. Given the built-in and growing market for driverless transportation, it is inevitable companies will try and step up to meet that need. Waymo and the car manufacturers exist to make money; this is not a charity project and they must see a chance to profit from driverless cars. They may fail of course, but the pace of progress suggests otherwise.

Finally, while the roads driverless cars use are public, this is not a government project in the same sense as the bullet train in California. Though the financial projections for the train say otherwise, no one with an ounce of sense expects the train to make money. It's almost certain it won't even cover its operating expenses when it's finished. In contrast, the development of driverless technology is being advanced by private profit seeking companies. In my opinion, that almost guarantees a different outcome than most government led projects.

Myob
Myob

I work at a self driving tech company, and Mish is right, this is inevitable, and a matter of timeline, but also the publicly stated timelines are a bit optimistic. Once you see how the sausage is made, you realize that making a self driving car which works in all conditions is a very hard problem. People thought early on that it shouldn't be too difficult, but at that time, we had no idea about what we didn't even know. Now, we have a better idea of where we're lacking.

Waymo is far ahead of everyone, but their systems have issues in rain (because reflections are really bad for lidar sensors, and rain confuses radar, and cameras have a harder time seeing). The cars are trained on clear roads, so snow also confuses them severely. I realize they have a snow clip in their video, but I guarantee you that they won't take passengers in snow. What they've got is a car that can drive safely in good weather. It will be years before it can drive safely in rain and snow.

Self driving cars today are not as good as a good human driver, but unlike the good human driver, they never lapse in attention, never look at their cell phone, and never tire. Their cameras don't see as well as people's eyes, but their other sensors make up for it in good weather.

Jojo
Jojo

Prostitutes will be using self-driving cars as their mobile hotel rooms!

Wagner_5
Wagner_5

It amazes me how one can write article about self driving tech. In the article mention Ford, Honda, Daimler, GM and what not, but not mention Tesla at all. Tell me which other car beside Tesla is capable of:

  1. exiting and merging highways and take into acount car in adjacent lane to share the road. Suggest highway lane changes in case you are stuck behind slower car.
  2. depict in center console number of lanes to the right and left by looking at lane markings
  3. recognize and depict in center console cars not only in front of you, but also to the right and left
  4. recognize and depict in center console semis, suvs, passenger cars, bicyclists or pedestrians
  5. allow to engage lane keeping (aka Tesla autosteer) not only on pre-mapped major higways, but pretty much anywhere where lane markings are present
  6. warn driver that he is entering construction zone

Ok, high-end Mercedes and Cadillacs can do some of those things, but only subset. And Mercedes are more expensive than Model 3.

P.S. I concede to Waymo which has better tech for city driving. However, I have never seen a Waymo car on Highway. Also, it is impossible for me or anyone else not working for Google to judge how good Waymo cars are at autonomy because google employees training/driving those cars don't post on public forums when car almost killed them (I believe Google would immediately fire them if they did that). On the other hand Tesla drivers use autopilot on highway and very often post on forums when mishaps do happen.

her_hpr
her_hpr

A couple of questions:

  1. price of a new vehicle.
  2. cost of maintenance of all those sensors and computers.
  3. who bears the liability when something goes pear shaped.
  4. 'solution' to the "trolley problem"
  5. weather . . .
thimk
thimk

well if self driving cars will reduce some of the 4000 fatalities on the road and perhaps eliminate many accidents will we see a reduction in auto insurance rates ? oh and what if someone "hijacks" the autonomous vehicle's software and commandeers the load. A lot of missing beer trucks.

Brother
Brother

We have yet to see one of these in heavy rush hour traffic or when traffic is 85mph in all lanes with heavy traffic.

Clintonstain
Clintonstain

How about we make this interesting Mish.

I’ll bet you $1,000 that “mass adoption by highway trucking” doesn’t happen by 2022.

Let’s see you put your money where your mouth is. Not a troll, just sick of the nonsense. How about we pay it off in silver.

mark0f0
mark0f0

One splash of water/ice/sleet/whatever and all those fancy sensors are useless.

Product liability will kill SDCs. Lawyers will sue everyone who ever touched a SDC or designed a part for it whenever there's an accident. Every company will have a 'smoking gun' or two in its files, and those will all be available to plaintiff's attorneys through discovery. Think of the General Aviation industry's litigation problem, but on steroids. Even manufacturers of trivial doohickeys on SDCs will find it cheaper to pay settlements than to go through discovery and the litigation defense process only to have claims defeated and be unable to recover costs.

Besides, does it really make sense to buy a million+ dollar SDC to replace a $20k/year driver? Lol. Do the math on that.

Snow_Dog
Snow_Dog

“At the beginning of the year Daimler announced a deal with Uber to introduce their self-driving cars on Uber’s ride-share platform in the coming years”

Now then, if only Uber could somehow figure out how to make itself profitable, you might have yourself something.

pgp
pgp

Great technology. Expect to see a lot of very smart but very stationary cars in a couple of decades when the oil and lithium runs out. Pedal powered self-driving bicycles will be more difficult.

Deter_Naturalist
Deter_Naturalist

Remember when there was a discussion of the algorithm that would be used by a SDC to decide what to hit at 55 mph, what it thinks is a group of pedestrians or an immovable concrete bridge support?

Would the "car" decide to kill the passenger rather than two or more "others?," if a collision was unavoidable?

Self-driving cars without the means to override the automation will have a heck of a steep incline of adoption to climb. I wouldn't ride in one.

Automated mining trucks and mining equipment? You bet. Automated combined harvesters for ag? Of course. Automated long-haul trucks sharing the road with cars? Perhaps. Fully-automated cars without steering or brakes from occupants? Very doubtful.

aqualech
aqualech

millions of new radar sources bathing all urban dwellers in radiation

Sechel
Sechel

I still think we'll see adoption self driving trucks before we see self driving cars. There's a safety argument to be made combining a driver with the self-driving vehicle. flees can drive in convoys creating aero-dynamic and fuel efficiency gains as the vehicles draft off each other.