Google, Detroit diverge on road map for self-driving cars

http://uk.reuters.com/article/2014/06/30/uk-google-detroit-idUKKBN0F50BV20140630

by kyrra • 12 years ago

69 81 12 years ago

uk.reuters.com

"Some industry observers have suggested that Google should pair up with Tesla, which is also developing self-driving technology and which shares Google's Silicon Valley mindset."

I wish the article had expanded more on this topic. Both companies could make significant gains by partnering and cutting out the old money from the automotive industry. Not that I'm saying I hope that happens, I'm neutral on that matter for now, but I'd like to see a lot more R&D done on autonomous driving tech in my lifetime.

rondon2• 12 years ago

It is much better to develop self driving technology in-house rather than risk being stuck licencing technology from Google in perpetuity. The software is the easy part. The regulatory and societal hurdles are the hard part and Google is already working on that for free.

ams61102• 12 years ago

A big selling point of Tesla cars is the performance, which is irrelevant if you are not driving. Teslas are sold as fun to drive, to people who like driving.

ryandvm2• 12 years ago

It seems obvious to me that industry outsiders are going to have a drastically different road map for self driving cars than the entrenched players.

Clearly the end game for self driving automobiles is far, far fewer cars in the world. In 25 years, relatively few people are going to be owning cars at all. You'll subscribe to a self-driving car service (or maybe your government will run it) and whenever you want to go somewhere, you'll enter your automotive needs into your mobile device and five minutes later the appropriate self-driving cycle/car/truck/van will pick you up in front of your house.

There won't be a need for every family to own 3 cars that spend 90% of the day turned off and taking up parking space. Of course Google and Detroit aren't going to see eye to eye on this...

6d0debc0712• 12 years ago

There may not be a significant reduction in cars if everyone wants to travel at the same time. Say morning and afternoon rush hours.

x1798DE5• 12 years ago

If you assume the point that people would prefer to use self-driving car services over having their own vehicle, then I imagine this wouldn't be as big a factor, since we already have a ton of unused vehicle capacity (with most people driving around with 1 or 2 people in a car that seats 4). You can imagine that modern car services with a significant customer base would be able to coordinate "carpool" type situations to significantly decrease spare capacity.

x1798DE5• 12 years ago

I'm really not sure that this is the case at all. Taxis already exist in basically every city in the US, as do vanpools, bus services and heavily subsidized public transportation, but we still have significant rates of car ownership. Self-driving cars do improve on the "human-driven-car" model because they have faster reaction times and the like, but the aspect of self-driving cars where you don't have to be in control of the vehicle isn't anything new.

It's certainly possible that cost was the major driver of these sorts of things, but my guess is that the convenience of having your own car ready at all times will remain paramount. It's also plausible to me that airbnb-like services renting out self-driving cars will be common, which would simultaneously alleviate the need to own a car and reduce the cost of car ownership.

I imagine the switch to primarily self-driving car services will have a greater impact in cities, where the density will be high enough that the latency between wanting a car and getting one will be smallest. If a car has to drive for 20 minutes just to pick you up and take you to the store to get some milk, you'll probably just want your own car.

abruzzi• 12 years ago

I've frequently heard the idea bounced around that self driving cars will eliminate the need for ownership. I'm glad you mentioned taxis, because it brings up something that has bugged me about this idea for a while. I live in a metro area of 200,000. This metro area has exactly zero taxis. (We do have a free ride services that is funded by the local city for getting the elderly and disabled around, which probably has six cars.) The three reasons are size density and wealth. As density goes down, mileage to get from point A to point B goes up, which means costs go up. It also means that cars are potentially occupied for longer periods, so you would likely need more cars to service the same population in denser areas. So ultimately taxis are not economically feasible here. So because you don't have to hire drivers, those costs go down and potentially lower the barrier to entry, but you have potentially more expensive vehicles, and higher maintenance costs that may somewhat offset the savings. So you could charge more for the service here, but being a poor city, would anyone here pay the extra costs? Enough to sustain the business model? Or will we have to suffer with significantly worse service--wait an hour to get picked up. I'm sure the business model would work in SF, NYC, Chicago, but here it will have to come down a lot in cost. Also (to the parent poster,) I just see the possibility that a migration like this could take place within 25 years just ludicrous.

bluedino• 12 years ago

> Taxis already exist in basically every city in the US

They exist, but in most towns nobody uses them at all.

duaneb• 12 years ago

Taxis are too expensive to use casually for most people.

slavik81• 12 years ago

Taxis are expensive and slow. There are relatively few of them, so you spend 20-40 minutes waiting for them to pick you up, and then pay them 5x-10x as much as it would cost you to drive yourself.

If self-driving cars were even half as common as regular cars, there would be a dozen parked on the street by your house. You wouldn't need to wait more than a few minutes, and cost could be lower than standard car ownership due to better utilization of the vehicles.

ack• 12 years ago

The idea of renting your own car out is already in practice (though not a self-driven one).

see http://www.snappcar.nl

Avshalom• 12 years ago

The only way the wait is going to be 5 minutes is if its already in your driveway.

Estragon• 12 years ago

In my experience, it's already less than 5 minutes for Lyft in SF.

infogulch2• 12 years ago

I wanted to take this further to get some utilization numbers.

1. In 2012, Americans spent 1 hour/day travelling (where work travel is about half of that). That's about 4% of their time. [1]

2. There were 253M registered vehicles in 2012 [2].

3. There were 313M people in the US in 2012 [3].

So a first-order approximation of vehicle utilization: (253M vehicles * 1 year) / (313M people * 4% time travelling * 1 year) = 2.2 trillion vehicle-hours / 110 billion travel hours = 3.4% utilization. So that "90% of the day turned off and taking up parking space" is more like 96%.

It would be interesting to take this further to expenditures. Consumers spent an average of $9,000 on transportation in 2012 [4]. If we increase vehicle utilization to, say, 21% from 3% using a subscription/micro-rental service like parent described, and assuming per-vehicle costs are constant, how much would the average consumer save on transportation costs? And how much would we save on parking space?

[1] http://www.bloomberg.com/visual-data/best-and-worst/most-tim...

[2] http://www.statista.com/statistics/183505/number-of-vehicles... See also for similar 2009 data: https://www.census.gov/compendia/statab/2012/tables/12s1096....

[3] http://www.census.gov/popclock/

[4] http://www.bls.gov/news.release/cesan.nr0.htm

Donzo• 12 years ago

I suppose one of the larger hurdles would be that during this 4% time that you calculated, a majority of the vehicles are in operation.

If "everybody" still has to go to and from work between the hours of 8-10 AM and 3-6 PM, pretty much everyone's going to want their own car.

hunvreus• 12 years ago

Probably even less utilization if you take into account public transportation, which probably make up for a significant portion of commute time in large urban areas.

jiggy2011• 12 years ago

Cars have more functions than purely transport.

They are great containers of stuff, you can leave tools, laptops , spare clothes etc in them without having to load and unload every time you want to go to work.

They're also status markets, so I suspect the rich will want to keep their own cars for this purpose which will lead to the middle class also being hesitant to give them up.

You also have logistical problems like what if I take an autonomous car for the day and decide I need to take a long cross country trip in it?

You also have special requirements for some cars, like perhaps I need access to a car with a lot of seats or I want to be able to carry a trailer and I can't guarantee that a car fit for this purpose will always be available.

josteink• 12 years ago

Them you do what people already do today with rentals: you check for availability and plan accordingly.

Slightly inconvenient compared to ownership, but much cheaper.

Some people will side with convenience, others with economics.

eftpotrm2• 12 years ago

Far fewer cars possibly, but a big increase in vehicle miles.

At present cars largely follow us round. If they're not doing this but are instead moving on to the next occupant, the odds are that that person won't be in the exact same location, and the next journey, and....

(Also bear in mind a huge amount of the demand clusters around start and finish times for offices and schools. The possible utilisation is significantly lower than one might think.)

If self-driving cars working on a taxi model rather than a personal chauffeur model become a thing, we're going to get used to seeing traffic jams consisting significantly of empty autonomous cars. That won't be especially popular.

toomuchtodo• 12 years ago

If self-driving cars working on a taxi model rather than a personal chauffeur model become a thing, we're going to get used to seeing traffic jams consisting significantly of empty autonomous cars. That won't be especially popular.

Autonomous cars that are "deadheading" (traveling empty to their next point) could be ushered into segregated high speed lanes/corridors (100+ mph), as no passenger means no safety issue.

fredophile2• 12 years ago

I'm pretty sure it would be easy to predict the spikes in usage. In the mornings people mostly want cars near their homes to take them to work. In the afternoon/evening they want cars near work to take them home.

Even without knowing this it's pretty obvious that if you have a large portion of users heading to one area (increasing the density of users in that area) they'll probably want to leave that area or move around it so you should have a higher density of cars in that area even if they aren't all being used.

_greim_• 12 years ago

Clearly the end game for self driving automobiles is far, far fewer cars in the world. [...] Of course Google and Detroit aren't going to see eye to eye on this.

Arguably, reducing the number of cars in existence at any given time doesn't reduce the number of cars sold. Rather, you'd have to reduce the rate at which they wear out and/or reduce peoples' reliance on transportation.

foobarian• 12 years ago

I'm skeptical about the ownership projections. It's very hard to beat the latency of owning a car that sits ready to go in the garage.

Owning a self-driving car could be interesting though. Perhaps you could let it participate in a ride provider while you're at work or sleeping, having it return well before you may need to use it. Also, it enables some interesting new activities such as you drive to a boat launch, take off in a canoe, and have the car drive downstream and wait for you.

higherpurpose• 12 years ago

This is why I think the major "car companies" of the future won't be the current "auto companies", but the current major tech companies. The "tech" part is going to become increasingly a bigger competitive advantage, while the "auto" part will be completely commoditized, and you'll see current car companies going bankrupt or being bought out by the major tech companies. I think Tesla is mostly part of the "tech companies" than the auto companies.

noir_lord• 12 years ago

while the "auto" part will be completely commoditized

To some extent this is already the case with some manufacturers, many models across a range are built on a consistent platform http://en.wikipedia.org/wiki/Automobile_platform

There is a point where physical improvements on platforms are well into diminishing returns and the bigger improvements are in software and control (the classic "software is eating the world" thing applies here).

I don't have a problem with this but I suspect the automotive industry might ;)

zenbowman• 12 years ago

Just to clarify, it seems that by "tech companies" you mean "software companies".

I think it is a little funny that a company like AirBnB would qualify as a tech company (even though I would classify it as a travel agency that uses technology) whereas GM would not (I would consider GM an out and out tech company).

But I agree that software will have more of an impact on vehicle sales performance in the future, I'm only objecting to the idea that software=tech.

mempko• 12 years ago

Yet, the car company who is bringing electric to the masses is the likes of Nissan.

Even how software is made via things like lean software development, scrum, and some agile practices are rooted in the Toyota production system.

Companies like Toyota in their very culture is lean, not bloated.

Even Google, despite all the diverse projects, only knows how to make money one way. Toyota group is far more diverse.

Also car companies in japan are heavily invested in robots. So there is plenty of disruptive tech happening there.

webmaven• 12 years ago

Interesting quote:

"Moreover, a study by consulting firm KPMG last year found that American consumers would trust brands like Google and Apple more for self-driving cars than they would automakers."

Which reminded me of this from a while back:

"Early concept research had made it clear that cueing the GM name resulted in a substantially lower quality perception and credibility, whereas cueing a Japanese name (such as Sony) did the reverse."

http://books.google.com/books?id=OLa_9LePJlYC&lpg=PA1946&pg=...

Pxtl3• 12 years ago

Too bad so many Asian brands seem hell-bent on destroying that good name with cheap garbage. Vaio laptops, anyone?

nodata• 12 years ago

Yeah give me an American Dell any day! Oh, wait...

jccooper• 12 years ago

Incumbents are correct that adding evolutionary features will allow maturation of technology and public acceptance. However, you cannot use that route to get to a completely autonomous vehicle. There will almost certainly be a red area once you get to the "car only needs human attention X% of the time" phase, where X will be somewhere between 60 and 1. Once the human is allowed to tune out, they're going to and won't be able to come back fast enough in many many circumstances. Heck, that happens today when X is 100.

Once you find X, you're going to have to stop and/or back off a little on automation. You can't just keep on pushing towards 0. I presume Google has already found the X point during their program.

Google is ultimately right to be aiming at fully auto if the goal is fully auto, not just "driver assistance." There's a big area where half measures just won't work. I wager it'll take quite a lot of work to get to that goal, but it at least seems achievable. And appropriate for an entity not already "in the business".

Pxtl3• 12 years ago

Exactly this. The moment you make an autonomous vehicle where some period of time can be done with no driver interaction, drivers will fall asleep. What do you do when it's time to get the driver back into the action, and they don't respond?

If expressway shoulders were consistently safe to pull over upon, that would be a good approach, but they're not.

Shivetya• 12 years ago

Google doesn't have the experience with liability that traditional automobile manufacturers are and I bet they tend to brush over items that made the other guys pause.

Its one thing to deliver ads and mail, a simple mistake or improper product usage by the end user doesn't hurt someone, but cars are a whole different ball game.

and don't underestimate that some people who get in one will try to cause problems.

x1798DE5• 12 years ago

Its one thing to deliver ads and mail, a simple mistake or improper product usage by the end user doesn't hurt someone, but cars are a whole different ball game.

Except for when they released Google Buzz and exposed a bunch of private data to the world, endangering some people who were relying on their trust in google to keep themselves hidden from abusive exes and/or various other dangerous sorts of person. Oopsie!

Gustomaximus• 12 years ago

Google doesn't have the experience with liability that traditional automobile manufacturers

Consider how Elon Musk is disrupting entire industries by refusing to think like this. IMO inexperience can be an advantage when competing with an entrenched industry. And experience can easily be bought where needed by a well funded company.

eftpotrm2• 12 years ago

Driving home from visiting a friend recently I had a lightbulb moment on self-driving cars; much as I really enjoy driving (heck, I've got a sports car), the second a mainstream car with an autonomous mode becomes remotely affordable for me I will buy it.

I confess I don't understand why the focus is so strongly on the taxi-replacement market; it's interesting, sure, but largely unviable outside metropolitan areas. It can't serve people who go shopping and want to use their cars as lockers during the shop, or people who leave sports gear in the car while they work then go to the gym later, or a number of other cases. Who pays for them? It'd be an expensive business to set up and you'd have regular arguments about taxi licenses that'd make the fuss about Uber look tiny. Who'd clean them? You bet people would do unsavoury things inside them. No; it's a small, awkward to serve.

Fit a 'self-drive' button to my next mainstream car though, and you have a killer product. You can keep it just like your normal car whenever you want, but when you don't - just press the button, program the navigation and do what you want. That night I'd have slept. This evening I'd probably have read on my way home from work. The technical challenges are no greater and it's much easier to set up.

That's a killer app, not an autonomous taxi. Your car, with an invisible chauffeur on demand. The first manufacturer to properly pull it off won't be able to keep up with demand.

fredophile2• 12 years ago

While taxis are one likely use for these cars I expect they'll also replace existing car share/co-op programs. I used to live in an urban area where I didn't need a car 99% of the time. I was in a co-op so I could book a car for a few hours the other 1% of the time. I don't see anything preventing someone from booking an autonomous car for a few hours while they get all of their shopping done. This type of use can also replace a second car for some families.

dragonwriter2• 12 years ago

I confess I don't understand why the focus is so strongly on the taxi-replacement market; it's interesting, sure, but largely unviable outside metropolitan areas.

Yeah, but metropolitan areas contain lots of people, and also fairly high concentrations of money, so serving them is particularly attractive.

It can't serve people who go shopping and want to use their cars as lockers during the shop, or people who leave sports gear in the car while they work then go to the gym later, or a number of other cases.

Well, sure, it doesn't serve the "use a vehicle as a stationary storage facility" role very well. OTOH, if it sufficiently reduces the per-use cost of using a vehicle as a vehicle compared to traditional ownership or taxi-with-a-driver, it can free up money for alternative, at-destination lockers.

Who pays for them?

Lots of possibilities, there: cities (perhaps particularly suburbus) as last (few) mile(s) connectors for mass transit systems. Or companies that want to disrupt the taxi market.

It'd be an expensive business to set up and you'd have regular arguments about taxi licenses that'd make the fuss about Uber look tiny.

So, maybe Google just does it directly. Its not like they don't have a huge pile of money to pay the startup costs.

Who'd clean them? You bet people would do unsavoury things inside them.

So, like a hotel, you do a user agreement that covers that, and have internal video monitoring (like most public transport) to identify the source of problems and assign responsibility.

Fit a 'self-drive' button to my next mainstream car though, and you have a killer product.

Yeah, that incremental upgrade is the more obvious direct market for consumers, and its easy to see why conservative players are focussing on it. Its also not surprising that Google's vision is bigger and less conservative.

sbierwagen• 12 years ago

You have enough money to own a car, which implies you're old.

Very few of my age cohort have a driver's license. Even if we had the time and money to get licensed, due to student loan debt we can't afford a car, or insurance, which is much more expensive for the young.

chris_mahan• 12 years ago

I'm old too. (45) and I have 2 cars in $30K-$40K range.

I would buy the self-driving car (2 of them) because I could read in traffic, or relax while listening to music, or watching a movie.

And if tired, on the way to Vegas from LA, push "autopilot button".

It's safer to let autopilot drive than to try to drive sleepy, trust me.

joelrunyon• 12 years ago

You can buy a solid car for less than 5k.

You don't need to spend 40k to get a vehicle that works.

Read this --> http://www.tropicalmba.com/entrepreneurmobile/

TwiztidK2• 12 years ago

I'm pretty young (22) and the overwhelming majority of my peers are licensed and usually have cars (disclosure: I don't have one myself). Although I do believe youth car ownership rates are much lower than in the past, there are a lot of places where it is near impossible to live without a car. The cars young people drive are usually cheap and old (potentially hand-me-downs from their parents or older siblings) and they only have liability insurance, but it is possible for a lot of people.

twotwotwo• 12 years ago

There's a big mapping problem that could maybe be solved by selling cars that have plenty of sensor hardware--can figure out curb positions to the inch and all that--but at first only use it for semi-autonomous features (lane assist, auto emergency braking, etc.). As people drive them, they're recording a ton of data that can be sent back to the mothership to build detailed maps will be useful to provide full-autonomous driving later.

Google might effectively pay for the data and field testing they're getting from early users by charging a lower price than you'd ordinarily expect for all that fancy hardware.

I don't know how all this will, in fact, play out; I'm just saying driving assistance and full autonomy don't have to be rival approaches.

stuaxo• 12 years ago

Just work with a car company from a different country that isn't so backwards.

I'm sure Volvo would be interested with their focus on safety.

lovemenot• 12 years ago

This discussion should probably start from commercial rather than passenger vehicles. In the latter category the driver is usually the owner as well as the cargo and the principle agent (my driving pleasure). There is an inherent distortion in the discussion around what the current drivers imagine they'd want.

Commercial vehicles live or die on their economics alone. Separating concerns in this way makes the issues clearer. Assuming they become established in the commercial market, producers can then address the consumer market. At that point, car drivers may rationally decide whether they'd prefer to pay the prevalent market premium for not using a self-driving vehicle.

Amazon might never deliver by drone, but the market for micro-deliveries by autonomous, wheeled vehicle could be huge.

ChuckMcM• 12 years ago

This reads a whole lot like an industry that has lost touch. If you read some of the stories out of SpaceX when they talked about Aerospace companies quoting 6 months and hundreds of thousands of dollars to develop a part, that SpaceX turned around and did in a few weeks, you realize that there is the status quo, and the folks who don't accept the status quo, and the people who try to enforce the status quo but have forgotten (or never knew) why it was that way in the first place.

Self driving cars are sexy and all but there is so much that could be done with a much simpler system.

The old 'personal rapid transit' ideas come to mind. For those too young to remember, that was a system where the cars would show up at the station when paged. The cars were on tracks but travelling either on a 'through' track (no stops) or pulling into the station when passengers enter/exit. Built using rubber wheels on a paved "track" such a system could be built more cost effectively than highways (no trucks) and with much better efficiency than metros (your car stops at your entry station and exit station only.). All the technology to build that exists today, and it would be massively more efficient than BART, but it isn't even being advocated.

iandanforth• 12 years ago

If you think Detroit is out of touch and slow moving I can hardly believe you're advocating getting hundreds of municiple governments involved.

gregpilling• 12 years ago

“We’d say, ‘Well you don’t really know that much. And we’re not going to put our name on a project like that because if something goes wrong, we have a lot more to lose.’” ----- I think that says a iot about the automaker point of view. They have experience with being sued for defects decades later, and they know the story of Audi and alleged "unintended acceleration"

pjspycha• 12 years ago

I just don't see computers being able to be better drivers then humans. Especially when dealing with something a programmer hasn't thought of or situations which are hard to predict.

I would definitely would prefer my taxi driver to be a person than a computer. (and honestly the trip would take twice as long in the city with a computer driver)

Google, Detroit diverge on road map for self-driving cars

Related Stories