Hacker News new | comments | ask | show | jobs | submit login

I think driverless cars could be the future of public transportation. Instead of big trains and buses with limited reach, we would have thousands of public vehicles that take you where you need to go when you need to be there. It eliminates the big barriers to public transport--the lack of personal space, timing, and reach.

No car? Just hop on the municipal/county dispatch website, request a car, pay your fee, and wait for the nearest open vehicle to bring itself to you. It can even coexist peacefully with private transportation.

How do you deal with the inherent inefficiency of having a separate vehicle for every passenger (or small group of passengers)?

Publicness is not the only benefit of public transportation. Trains and buses are incredibly efficient per passenger mile. They also take up a lot less space than cars, and last longer and cost less per passenger to manufacture. Making a car driverless doesn't solve these problems.

The de facto standard in this country is one dedicated car per person, which spends 90% of its time sitting in a parking space. Communally shared cars will be massively more efficient than that, and, unlike buses and trains, will actually replace dedicated vehicles for a majority — some day 100% — of users.

The public transportation the public uses beats the one they don't.

Not to mention that communal vehicles would make electric cars suitable for any length of trip— Forget about exchanging batteries, just swap cars at the charging station.

Communally shared cars would still have the problem that a large number of those personal cars are spending the same 90% of their time sitting in a parking space. Everyone's leaving for work at 8, leaving for lunch at noon and leaving for home at 5.

You'd still need roughly the same number of 'communal' cars during those peak times. Though people may be more likely to car-pool.

Even the trivial variation of most US 'flex time' starts wouldn't help much. The difference between 8 and 9 isn't great enough for very many cars to make it from a business park at 8 back out to pick up another passenger and then back to another employer before 9.

What might really help, is 'communal' cars augmenting the US's sparse transit system. e.g. The communal car takes you to the train/bus stop. A communal bus departs as soon as it can. The communal car takes you to your final destination.

That way the cars are doing shorter round-trips, serving more people (necessitating fewer cars) and people are more apt to use the public transport because they aren't worried about the current big problems of US transit: 1. parking at the transit stop 2. how to get from the transit destination to where they really want to go 3. what happens if you miss the train/bus time (the car can just complete the trip)

That's true currently— but I think you're neglecting the possibility that widespread use of robotic cars could effectively eliminate rush hour. If that hour and a half commute starts taking the twenty minutes it should, two hours is time for a bunch of round trips.

Of course there will be some limit to capacity, as with any public transportation. But having to wait a bit for a car isn't going to put off commuters who are used to spending hours a day waiting in traffic.

That's not even mentioning you can actually do useful stuff sitting as a passenger during the commute.

"The de facto standard in this country"

Which country? It's a global problem (and this is a global forum). Kudos to Nevada for having the presence and foresight to pass this legislation. It opens up a world of possibilities for future transport options, and whilst it will take a while for the public to catch up and accept both the technology, and the transport changes it enables, this is a huge step in the right direction.

What country is Nevada in?

I'm not blind to the existence of transportation issues in other countries, but insufficient (read "nonexistent") public transportation is often noted as a predominantly (while not uniquely) American problem. Certainly we represent the archetype of the "one car per person" catastrophe.

The emergence of communal, computer-controlled electric vehicles could turn that around completely, so I don't think it's too out of line to be talking primarily about America at this stage.

"What country is Nevada in?"

Spain for one. I know which Nevada the legislation was passed in, but your comment made the implicit assumption that everyone reading it is in the US.

"insufficient (read "nonexistent") public transportation is often noted as a predominantly (while not uniquely) American problem"

As is egocentrism.

What are you talking about? There's no Nevada in Spain, GA. (Wait a minute... Oh. Gotcha.)

Seriously, though. "This country", as in, the country where I live, which is coincidentally the same one under discussion. You can relax a little with the righteous indignation.

Theoretically, driverless cars could be programmed to behave in a collective fashion. Yes, a train is incredibly efficient when 300 people need to go from the same place to a different place. But in reality, 300 people are coming from 300 places and going 300 places. Driverless cars could calculate partial carpools or transfers to more efficient vehicles like trains depending on the needs of the commuters. A "smart car" could drop the commuter(s) off at the train station, and another car could pick the commuter(s) up for the final leg of their trip.


A few weeks ago, I left my house for the grocery 2 miles away. I followed a car from my subdivision to the grocery. After shopping, I followed the same car back from the grocery to my subdivision.

Of course, this points out a failure in public engineering, not the need for automated cars. (As much as I'd love to have them)

Urban sprawl is a gigantic problem that we need to address, and soon. It not only affects transportation - there's general maintenance cost, social effects, land use, etc.

A car can't leave the station traveling west at 220 mph arriving at its destination in 30 minutes. Driverless cars will be great but we also need fast, efficient ways to move people too. I want mine to pick me up at the station when my train arrives.

More likely, driverless cars will be programmed to make really, really long commutes.

How far would you be willing to commute right now? OK, what if you could sleep or play videogames en route?

I commute about 1:15:00 each way, each day. I carpool, so I can read my iPad for two hours a day. It certainly helps with the drive time.

But you can do the same on a train or bus already. I think part of why people don't use them as much is because of the inconvenience of working around their schedule, and first and last mile commutes (because the train doesn't go to your office). Networked, driverless cars could play a part in the larger public transport system to alleviate these two points of pain of commuters.

Rush hour crowding on public transport is a big negative. If we embraced more flexible work hours, it would bring about its own efficiencies.

A country like the US, especially with vast suburban areas, it isn't necessarily a strict dichotomy between public transport and driverless cars. We can have both.

Especially in the west coast, American cities were built to be navigated by cars. Even today, the very fact that carpool lanes in most parts of California are designated for 2 people/car, not 3 implies we already have separate vehicles for each passenger. If we can take advantage of the road infrastructure and cut down on the need to own a car, it will be a huge benefit to lots of people. See how popular ZipCar is, precisely because of this sort of benefit.

Google Transit already offers you choices when you want to go from A to B. It could do the same thing with driverless cars. Want to take one the whole way? That'll be 60 minutes and cost $12.50. Or you can can take the driverless car to the train, which will take you the rest of the way. $7.50, and takes 1h20 minutes.

I know someone who spent some time in one of Japan's smaller prefectures. He hated how all the trains stopped promptly at midnight, making late night things impossible. Something like this would let you take a train there and an automatic car back.

This will definitely benefit bars. I would love to see drunk driving become a thing of the past at some point in my lifetime.

It wouldn't necessarily have to be the only public transit option. In New York, I can get to work by train much faster than by car - but if I need to go to, say, Ikea, it's two trains and a bus. A cheap, public car would be great.

I agree completely. I don't necessary think it's a bad idea, I just don't think it's a good idea "instead of big trains and buses."

I think of this more as dealing with the present (and unpleasant) reality. I prefer buses and trains, but we haven't found a way to get people to fund and use them enough to make them viable outside big cities.

A lot of the gridlock in my city comes from people doing stupid things that force other drivers to slow down to avoid a wreck. Automated cars would reduce the impact of bad drivers by removing them from the equation.

I've always thought it would be neat to combine the public transport with the public car idea. Have parking lots full of public cars available at the station stops (and really big ones at the end of the subway line).

You'll have to be able to keep one of these public cars at your place overnight for it to work (unless it drives itself home empty!).

More accurately, trains and buses can be incredibly efficient per passenger mile.

Anecdotes are not data, but our local buses (city of ~50k) rarely have more than one rider, and a plurality of those I see have zero riders. In our case, these autonomous cars would be a much more efficient solution.

"More efficient than a bus system nobody uses" is a pretty low bar. In your city's case, shutting down the bus system and having everyone currently riding the bus take (potentially subsidized) taxis would be way more efficient, both environmentally and financially.

There isn't that much inherit inefficiency; if the cars are smart they can drive in a train configuration, imagine this without the mechanical linkages:


But people don't use them that much! In theory trains and buses are more effecient, but since they aren't used that much, that means people don't care about that.

Or you rent your car out to the grid rather than park it. Like air-bnb for cars.

See getaround.com. They are doing something like this already.

getaround.com looks interesting. Shame they require a facebook account.

It's the future of "transportation", period. People will look back to the time when humans drove transportation and laugh.

Eric Schmidt has referred to it as "solving the bug that cars were invented before computers." It seems so obvious now!

And in the even farther future, people will look back on the time when it was legal to have a human driven car the same way we look back on safety standards in the early 1900's

I think it's only a matter of time (maybe 10-20 years, who knows) until one of the larger urban areas with huge traffic problems, such as Manhattan, London or Beijing mandates these sorts of systems in congested areas. It really seems like it could be the only viable solution to traffic problems in cities whose roads just can't handle the loads, and it seems like a logical next step after congestion pricing like London already does.

I'll be honest, I'm waiting for commercially available vehicles so I can be the first to order a fleet of them and do just that :-)

This would be a big boon to any of the cities around me (like Athens [GA]). They have limited street parking, so the fees are big and the times are short. This would let you go to where you're going and send your car over to a high-density parking garage across town to wait for you.

Do a lot of people avoid public transport just because of the lack of personal space?

Every person I know who has talked about public transport says the other people (especially bums napping and skeevy people) are the big thing that limits the amount of traveling they do on rails and buses. Reach and timing take up the next two slots.

I can definitely feel 'trapped' when people invade my personal space, especially if they smell or have a cough. Those experiences are horrible for the extended amounts of time required for a commute. I still ride, but driving really isn't an option.

Sometimes I wonder why public transport doesn't raise the prices during rush hour to create an equilibrium where the trains don't become sardine cans. I guess economically and somewhat ethically it might be harmful.

I pay $10 to get to work and back in rush hour. Parking is $12 for the day, plus gas and wear and tear. I would gladly pay some extra cash if it made the trains emptier, cleaner, and more reliable.

Do you want all those people to clog up the streets in their cars?

Lots of cities have problems with their transportation capacities, but simply shuffling people around isn't going to fix that.

Maybe a first class area for daily commute

Some of the trains used for commuting where I live (southern sweden/denmark) do have a first class area (which is nearly always empty…).

I find dislike of public transit strange. I have a strong dislike of cars - you have to pay for the car, you have to pay for insurance, you have to pay to fix it when it randomly breaks, you have to pay for gas, it lives outside your house on the street (so who knows what could happen to it late at night)... I could go on.

My ideal transportation arrangement is a subway pass in my wallet and nothing else.

Which, in the USA, pretty much limits you to New York, Chicago, and Boston, it seems. Not that those wouldn't be my ideal places to live anyways, but it's still frustrating.

Public transit-only travel is workable in Philadelphia and Washington, D.C. as well. Although in Philadelphia, you would have to take the buses quite often to compensate for the gaps in the subway system.

Personally I avoid it because it's not very reliable(in Houston at least. It was pretty reliable in Germany) and it's not that effective at getting me where I want to go. There's still huge sprawls of land that are actively used but not connected to the public transport routes.

I'm originally from Dallas; pubtrans there is a joke.

This already exists in Austin, TX and a few other cities around the world. http://www.car2go.com has Smart cars scattered around the city. You grab the nearest one (or book ahead with your smartphone), swipe your RFID card, drive as long as you'd like, and then leave it in any public parking spot.

Edit: Minus the driverless bit obviously. :)

The driverless bit is what makes this work. If I just hopped off the high speed rail from Atlanta to Austin (I'm optimistic!), I don't want to have to walk to the nearest Smart Car center. I want the car to come to me. :)

Then it doesn't exist.

Driverless is key.

I wonder if human-driven car will become minority of vehicles in next 30 years. If the technology will be affordable and good enough, this seems like a probable scenario. Also it could be possible that driverless cars will reduce traffic jams by introducing some better routing algorithms.

They'd reduce traffic jams simply by behaving consistently over time. A very large number of traffic problems are caused by distracted/daydreaming drivers, aggressive drivers, overly cautious drivers over-braking, etc.

Consider the situation of lane-merging. Studies have shown a zipper-merge at the end of the disappearing lane to be the most efficient method. Yet human drivers create huge traffic snarls in these situations based almost entirely on their emotional reaction to what they perceive as 'fairness', which hinges on a thoroughly baseless concept of 'winning' in traffic.

Introduce driver-less cars and --with no other changes to traffic patterns at all-- the rising incidence of proper zipper-merging will reduce the frequency and severity of traffic problems.

Big trains and buses still move a lot more people than one small car. If everyone hopped into their driverless car, we'd be back to where we are now -- gridlock :)

In theory driverless cars could end up having algorithmic optimizations and use networked information such as to use overall road capacity much more efficiently.

Then hackers can take control and direct all traffic to participating gambling establishments.

> In theory driverless cars could end up having algorithmic optimizations and use networked information such as to use overall road capacity much more efficiently.

A company like Google isn't smart enough to drive off about.com from the front page of almost all of their SERPS, and do you actually expect them to come up with an "algorithmic optimization" for driverless-cars in cities like Cairo or Istanbul?

Traffic routing is a much simpler problem than deciding which webpages are relevant to which search terms. A bunch of CS undergrads could probably come up with a prototype system.

... And it's going to be around three years, until everyone takes them for granted and bitch about every minor niggle. :)

And instead of 50-100 people on 1 vehicle, it's 1. Not really eco-friendly, and not cheap either.

These are also known as "taxis".

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact