squirtdad

humor....but risks of government dictating all about AV would a concern for lots of people

tyrion

They can't make enough AV cars to achieve a "domination of traffic" in 7 years. Not even close.

long john

The cars will see us as squirrels and speed up.

mr_bill

The rules for horseless carriages are exactly the same as the rules for carriages. (Carriages that turn into pumpkins at midnight the exception that proves the rule.)

-mr. bill

Ninety5rpm

Put me in the category of people who think AVs are going to very much change how people use cars.

It will go something like this.

1. AVs will be initially available only via hailing services (Uber/Lyft/taxis)
2. Hailing services with AVs will initially be limited to certain geographic areas.
3. People will initially try them as a novelty.
4. Prices for rides will be significantly lower than for human-driven rides.
5. Competition will be fierce.
6. Areas where AV taxis are available will grow rapidly.
7. People will figure out hailing a ride door to door at an inexpensive rate will make a lot of sense.
8. The high popularity of AV taxis will make efficient/convenient pooling possible for the first time.
9. Pooled AV taxi rates will be comparable to public transport fees.
10. Public transport will be virtually abandoned in areas where AV taxis operate.
11. People will start using AV services so much, they'll stop using their personal cars, and wish to sell them.
12. Pooled use of AV taxis will greatly reduce traffic congestion; most AV taxis will really be mini buses.
13. The geographic areas (radius) in which AVs can operate will expand, everywhere, exponentially (they will grow until they overlap and connect).
14. Rental companies will offer AVs for people who need access to the same car for longer periods.
15. There will be a glut of used human-driven cars available for sale. It will be a problem.

Ninety5rpm

Sure they can. One AV minibus can do the work of transporting several dozen people in personally-driven vehicles on a given weekday morning during the commute hours of 5am to 11am.

tyrion

That can happen in densely populated areas, but not in most of the world. And it would only work in densely populated areas if passengers' commutes were scheduled to keep the minibus busy for 6 hours. Not gonna happen soon.

Ninety5rpm

For pooling services to be efficient and convenient, a certain percentage of travelers within a given area have to be opting to use pooling services. And the lower the density, the higher that percentage has to be, obviously.

So, yes, initially efficient pooling will occur in high density areas. But as AV taxis grow in popularity, and the percentages using them keep growing, pooling will become viable in lower and lower density areas, until they will be viable everywhere there is significant traffic, where the AV taxis will dominate. Remember, in high density areas human driven taxis are already arguably dominating. The number of cars driving around with Lyft or Uber stickers in San Francisco, for example, is staggering. And this is before the big fare drops due to autonomousness and pooling.

Think about how much more hailing services are being used now simply due to the relatively modest fare drops and app hailing convenience offered by Uber and Lyft compared to traditional taxi services. When those fares drop by half or more, and half again (at least) when one agrees to be pooled, it simply won't make economic sense to drive your own car for most trips.

I think this will happen quite rapidly, within a few years.

tyrion

No it won't. The lower the population density the less attractive pooling becomes. It becomes more time consuming and more schedule restricting. People are willing to pay for the time saving and schedule freedom of owning their own car. That won't change in the near future.

Ninety5rpm

My point is if (say) 1% use pooling in an area with 20,000/sq mile (like San Francisco) that pooling is going to be just as efficient and convenient there as it is with 10% using pooling in an area with 2,000/sq mile (like Boise). Note: these number are just wild @ss guesses used to illustrate a point.

It's easy to look at the current percentage of pooling use in SF and say that's never going to work in Boise, but that's assuming the same percentage will be willing to use pooling in Boise as are using pooling in SF. In high density, even very low participation rates can work okay. In lower densities, that low participation rate isn't going to work. Of course.

But with (much) cheaper ride fares, I really think the participation rates will be much higher, making pooling viable in much lower density areas.

noisebeam

https://www.azcentral.com/story/mone...ar/1078466001/

Ninety5rpm

Way behind Waymo, Apple is never-the-less up to something too.https://www.bloomberg.com/news/artic...%20T%26L%20ALL

This is the free market at work. One of these approaches will work best, and we will all benefit greatly from the valiant efforts being made to find out which one that is.

And one of those benefits will be cycling safety: MUCH FEWER cyclists overlooked and hit.

Ninety5rpm

There is no way they are going to hire people to drive or babysit all those vans. Holy *****. That means a driverless fleet in operation sometime this year, two years ahead of the 2020 prediction I made... last week!

Ninety5rpm

OMG. I can't believe somebody is doing this. @I-Like-To-Bike is not the only one concerned about AVs getting into situations they can't handle. So this approach is to have drivers on the ready to intervene... remotely!
https://siliconangle.com/blog/2018/0...innovationday/

02Giant 

Then your connection with this thread must be recent.

It has been proclaimed that it will happen virtually overnight.

Ninety5rpm

On the order of smart phone adoption, perhaps faster.

And, to be clear, by "it" I mean autonomous vehicles are more predominant than human driven motor vehicles in traffic in US metro areas.

jefnvk

You mean how like ten years ago, every EV had its own plug, but now they are more or less standardized?

squirtdad

Apples/oranges.

A power plug is very simple thing. The specification for a standard plug can be done on a page or two of paper. Physcial specifications, Electrical specification, material specifications, some simple electronic specifications, possible some emedded software specifications

Sharing data between different automation applications/systems is an entirely different thing altogether.

there is no standard system, each use a proprietary mix of software, hardware, sensors (lidar, radar, gps, etc). The software approach to the solution and sensor inputs (hundreds of millions of lines of code) will be very different as will learning and sharing model.

there is business advantage to being better at this than your competition, so there is not motivation for waymo to share with GM.

Indeed even the input for sharing could be proprietary in that it would tell one company something about how another companies system works.

will require goverment intervention to happen and a lot of rework to existing systems..... not gonna happen any time soon

350htrr

Why...? Look at Betamax and VHS, it all got worked out, the worst system won...

02Giant 

First, I believe the speed and share of the transportation system proposed by some here, is decades away.

What you suggest will change by acquisitions within the market. If, the market starts to prove it is viable.

jefnvk

Nowhere near as difficult as you make it seem. The FAA is already implementing such a system for air traffic, if you want a more akin example.

Creating data wrappers to exchange mutually beneficial information such as location, destination, route and nearby items of interest to broadcast in some sort of ad-hoc manner doesn't require any devulging of trade secrets. I could easily point to the standardized OBD communications, put in place for emissions regulations, as another such specific example. It outputs information in a standard manner, regardless of different OEMs not even having the same in-car communication networks in many cases

Ninety5rpm

I agree with this, but , again, the data and experience each company is gathering and taking into account in their cars is way more than any one person can process in a lifetime.

Ninety5rpm

Oh, shti. I agree with this too. Good point.

What probably won't be shared is how each car company's cars deal with a given situation. But information about situations can certainly be shared.

And it doesn't even have to be electronic data that is shared; just knowledge of problematic scenarios in general. Case in point: Tesla fails to see white truck against sun and crashes into it. Guess what just got added to Waymo's and Apple's simulation and test scenarios?

I-Like-To-Bike

I dunno, do you you know? What got added? What got solved?

genec

Yeah... DVRs. Who uses tape anymore?