Self-driving car discussion catch-all

I am currently looking at replacing my 2004 Mini Cooper S and suspect that my next purchase will likely be the last human piloted, petroleum powered vehicle in my lifetime.

One thing I do wonder about with self driving cars is how well they will do on dirt roads. Or are we going to have to go with some kind of dual use car that has to be self driving on paved roads and manual on dirt roads/off-road.

I imagine dirt/gravel roads that are fully mapped won't be an issue.

For stuff like farm roads and the like there almost certainly won't be a steering wheel and pedals since it's likely that many people won't have any experience driving manually. I can foresee an interface on the vehicle computer where maybe you draw a line on the satellite map view with your finger and the car scans the ground and follows it. Or maybe you tap the destination and the vehicle launches a drone and maps the terrain choosing the best route. I think there will probably be a number of differing methods, with the most intuitive winning out in the end.

Isn't that why most of the cars are using lidar or cameras or whatever? They don't need GPS, they can see where the road is

They do still use GPS for knowing which road they're on, where to turn, the best route to take, etc. When an autonomous car comes to a fork in a mapped road, it already knows which way to turn. With back roads, that data isn't available, so that's where the user input is required.

The question is how will the vehicle acquire the user input? Traditional steering wheel, joystick, touch screen, or something else completely?

Automated race cars are already able to pick better lines with better speed control through turns than any human. I would imagine that the same kind of terrain recognition used in that sort of thing would make it possible for cars to master dirt/off road driving far quicker and to a much higher degree of mastery than humans.

Serengeti wrote:

They do still use GPS for knowing which road they're on, where to turn, the best route to take, etc. When an autonomous car comes to a fork in a mapped road, it already knows which way to turn. With back roads, that data isn't available, so that's where the user input is required.

The question is how will the vehicle acquire the user input? Traditional steering wheel, joystick, touch screen, or something else completely?

More to the point, this will be an incredibly short-lived problem.

Once you put automated cars on the road, every little dirt whatever that a car could possibly want to drive on will get "mapped", because there is an incentive to do so. It'll be part of the job of government transportation agencies or what-have-you, call them to come out there and drive their special "mapping" car over the path to plot it out, or something like that, and then submit the data for car GPS updates.

It would probably never happen, but it would be super cool if the USGS could set up a central mapping database and standard protocols that all self-driving cars could use to keep it up-to-date.

As it is, individual companies will probably never share that data because it gives them a competitive advantage or some other nonsense reason.

Paleocon wrote:
ActualDragon wrote:

Both of Paleocon's points are good. A lot of us in transportation are really hoping that automation helps us get away from car ownership, particularly in cities. The actual miles traveled are projected to increase by some experts for the reasons that you're citing - if we eliminate the pain points of being in the car, people will want to be in the car more. But if we decrease the number of vehicles and can take advantage of automation to get operational efficiency out of the existing network that we can only dream of now, we're looking at a pretty bright picture for mobility.

Though I do worry that the Ubers of the world will take this over and turn it into a money-grabbing capitalistic mess, instead of it operating more like public transit. If this tech is made accessible to everybody it could solve so many problems that our cities are facing. This is the backbone of the Smart Cities movement and also why it's housed under the US Dept. of Transportation despite involving many other disciplines.

It will definitely change how we use space. Parking lots/decks in prime business areas will largely disappear or consolidate around satellite parking megastructures (like the airport model). Traffic signalling will largely be invisible to vehicle occupants and pedestrians alike. Curbside parking will mostly disappear. I suspect that the fluid dynamics of the "loading zone" will become a lot more important than current considerations for parking.

I hope it all adds up to greater pedestrian utilization of space.

Depends on how long we remain in a transition state. That's likely to be ten years or more. And while a fully automated system will allow for much more capacity, there's still the issue of bandwidth. Everyone talking about how much more they were going to drive just got my wheels spinning a bit. So to speak.

BadKen wrote:

It would probably never happen, but it would be super cool if the USGS could set up a central mapping database and standard protocols that all self-driving cars could use to keep it up-to-date.

It's already happened. It turns out, though, that the data isn't nearly enough for mapping, let alone self-driving cars. Fortunately, there's many alternatives, and they've been improving by leaps and bounds for years.

As it is, individual companies will probably never share that data because it gives them a competitive advantage or some other nonsense reason.

It's already being done. The real cost in accurate mapping is maintenance - Google's operation, for example, costs tens of millions of dollars a year. Sharing those maintenance costs is much easier for everyone, and results in a more accurate, more up-to-date map.

U.S. House unanimously approves sweeping self-driving car measure

The U.S. House on Wednesday unanimously approved a sweeping proposal to speed the deployment of self-driving cars without human controls by putting federal regulators in the driver’s seat and barring states from blocking autonomous vehicles.

The House measure, the first significant federal legislation aimed at speeding self-driving cars to market, would allow automakers to obtain exemptions to deploy up to 25,000 vehicles without meeting existing auto safety standards in the first year. The cap would rise over three years to 100,000 vehicles annually.

Representative Doris Matsui said the bill “puts us on a path towards innovation which, up until recently, seemed unimaginable.”

Automakers, business groups, and advocates for the blind praised the House measure. But one consumer group said the House bill did not do enough to ensure self-driving cars would be safe.

Under the bill, manufacturers seeking exemptions must demonstrate self-driving cars are at least as safe as existing vehicles. States could still set rules on registration, licensing, liability, insurance and safety inspections, but not performance standards.

Automakers would have to submit safety assessment reports to regulators, but the bill would not require pre-market approval of advanced vehicle technologies. The measure now goes to the Senate, where a bipartisan group of lawmakers has been working on similar legislation.

Automakers and technology companies, including General Motors Co and Alphabet Inc’s self-driving unit Waymo, hope to begin deploying vehicles around 2020. They have been pushing for new federal rules making it easier to deploy self-driving technology, but some consumer groups have sought additional safeguards.

Current federal rules bar self-driving cars without human controls on U.S. roads. States have issued a variety of different rules in the absence of clear federal guidance, and automakers have complained that California’s rules are too restrictive.

U.S. senators might circulate their draft legislation this week. One sticking point is how to handle commercial self-driving trucks, which are not included in the House measure. The Senate version may also soften the provisions preempting state rules.

Volkswagen AG (VOWG_p.DE) and other automakers have been lobbying Congress to act, often bringing test vehicles to Capitol Hill so lawmakers can test out driverless cars.

Advocates hope self-driving cars can help reduce U.S. road deaths, which rose 7.7 percent in 2015, the highest annual jump since 1966. The U.S. National Highway Traffic Safety Administration said in a 2014 study that U.S. traffic crashes cost society $836 billion a year in economic loss, with human error behind 94 percent of crashes.

Consumer advocates want to give the National Highway Traffic Safety Administration quicker access to crash data and more funding to oversee self-driving cars.

“The autonomous vehicle bill just passed by the House leaves a wild west without adequate safety protections for consumers. It pre-empts any state safety standards, but there are none at the national level,” the Consumer Watchdog group said in a statement.

The policy group Transportation for America said cities are worried the House “legislation will preempt local authorities from managing their own streets and fail to give local leaders the confidence that manufacturers and operators will be aware of and follow local laws and regulations.”

On Tuesday, Reuters reported that U.S. Transportation Secretary Elaine Chao will unveil revised self-driving guidelines next Tuesday in Ann Arbor, Michigan, citing sources, and the department confirmed plans to unveil the new guidelines next week. The House bill would require compliance with the guidelines.

GM said in a statement that “while more work is needed,” the House measure is “good progress toward a law that will facilitate realization of the safety, mobility, and environmental benefits of self-driving vehicles.”

The House bill would also require automakers to add a driver alert to check rear seating in an effort to prevent children from being left behind, and to consider performance standards for headlights.

I'm shocked that the House voted so strongly for something that will clearly put people out of jobs, not to mention increase the role of Federal regulation in a market. Are we no longer in the darkest timeline?

I think Detroit just wants to sell more cars.

Imagine a point where we reach critical mass and self driving cars become mandatory or something. They can sell a brand new car to everyone!

Yeah, which do you think has a bigger lobby, auto makers or truckers?

Also, don't forget, the number of trucks on the roads is limited by the number of truckers that can drive them. If you don't need truckers, you can have a nearly unlimited number of trucks on the road. Truckers lose, truck makers win big.

Making truckers unemployed also means you can win elections 10 years later with ridiculous promises of bringing back trucker jobs.

Shadout wrote:

Making truckers unemployed also means you can win elections 10 years later with ridiculous promises of bringing back trucker jobs.

Plus, you've got a built-in market for your trucker hats with laughably simplistic political slogans on them.

Shadout wrote:

Making truckers unemployed also means you can win elections 10 years later with ridiculous promises of bringing back trucker jobs.

Oh yeah, I see it now. Faith in humanity restored.

I can see the slogan now, "Truck America".

"Make Truckistan Truck Again! TRUCKS!"

...

I may have lived too long in Texas.

Glory to Arstruckzka!

Some pizza company is testing out self driving cars. You have to sign up for the service though.

Mixolyde wrote:
Shadout wrote:

Making truckers unemployed also means you can win elections 10 years later with ridiculous promises of bringing back trucker jobs.

Oh yeah, I see it now. Faith in humanity restored.

And blame immigrants for taking the few low paying and dangerous trucking lobs that remain.

Have we discussed how self driving cars will affect vacations and vacationing? I'm currently at DisneyWorld and staying at the Polynesian resort, and a large reason of that is the convienve of being able to get on the monorail and go directly to the parks.

But. If self driving cars were super common would I do so? I'm not sure. Taking a 20 minute self-driving car from an off site hotel to the park wouldn't be that different, time wise, then monorailing it and it would actually be easier logistics wise.

Cheaper vacations as people won't care as much if they are super close to the object of the vacation, be it beach or theme park or whatever?

jrralls wrote:

Have we discussed how self driving cars will affect vacations and vacationing? I'm currently at DisneyWorld and staying at the Polynesian resort, and a large reason of that is the convienve of being able to get on the monorail and go directly to the parks.

But. If self driving cars were super common would I do so? I'm not sure. Taking a 20 minute self-driving car from an off site hotel to the park wouldn't be that different, time wise, then monorailing it and it would actually be easier logistics wise.

Cheaper vacations as people won't care as much if they are super close to the object of the vacation, be it beach or theme park or whatever?

You're not wrong, but it's statistically insignificant. Apply the same logic to where you work and where you live. You know, where you spend the other 51 weeks out of the year.

But to your point, the largest impact to vacationing is the elimination of the rental car industry. That's huge.

Jonman wrote:

But to your point, the largest impact to vacationing is the elimination of the rental car industry. That's huge.

Is it? Or are you just going to rent a self-driving car?

Jonman wrote:
jrralls wrote:

Have we discussed how self driving cars will affect vacations and vacationing? I'm currently at DisneyWorld and staying at the Polynesian resort, and a large reason of that is the convienve of being able to get on the monorail and go directly to the parks.

But. If self driving cars were super common would I do so? I'm not sure. Taking a 20 minute self-driving car from an off site hotel to the park wouldn't be that different, time wise, then monorailing it and it would actually be easier logistics wise.

Cheaper vacations as people won't care as much if they are super close to the object of the vacation, be it beach or theme park or whatever?

You're not wrong, but it's statistically insignificant. Apply the same logic to where you work and where you live. You know, where you spend the other 51 weeks out of the year.

Edit: Moondragon-hausered

But to your point, the largest impact to vacationing is the elimination of the rental car industry. That's huge.

But would the rental car industry be eliminated? Where would jrralls get that self-driving car from in Florida? In fact, until we (hopefully) eliminate human driving, I can see rental car agencies renting self-driving cars at a premium.

And to his point, really how would the situation be any different from renting a car today? Unless you start hammering beers at 9am (I'm not judging, sounds like kind of a good vacation to me).

MoonDragon wrote:
Jonman wrote:

But to your point, the largest impact to vacationing is the elimination of the rental car industry. That's huge.

Is it? Or are you just going to rent a self-driving car?

You're talking about a taxi service, not a rental service.

Same reason car ownership will go the way of the dodo.