The Coming Zombie Robot Driving Apocalypse of You

Generali

antrobus wrote: »

The point I was making that, for example, in 2012 there were 1,754 road deaths, and 420 of those were pedestrians, about a quarter of whom were drunk at the time. With the best will in the world you're therefore never likely to be able to reduce the number to zero.

I agree: you can't save people from themselves.

If you can reduce deaths from 1,754 to ~100 plus a few faults and mechanical errors then that's a pretty good thing. Especially if you can use roads more efficiently and reliably at the same time.

antrobus

Generali wrote: »

I agree: you can't save people from themselves.

If you can reduce deaths from 1,754 to ~100 plus a few faults and mechanical errors then that's a pretty good thing. Especially if you can use roads more efficiently and reliably at the same time.

That's kinda of the reduction that would be feasible. You'd tend to expect that 420 annual pedestrian deaths could be cut to 100-200 a year given that these robot cars should have a faster reaction time, and be able to launch the emergency stop protocol within milliseconds.

globalds

The cars for individuals may be a huge market.

But just like those paying for the first mobiles were companies buying them for key staff.
Then the first users of this tech will be businesses.

Somewhere ..And I would be inclined to think there are lots of somewhere's ,where a driver isn't needed to unload , where a delivery takes longer than a working day , where a vehicle only returns to drop a driver back. Where time is available between the going and coming back for a vehicle to do another task..

An extra £30,000 sounds a lot , until you stick it on the cost of a new artic. Especially if it allows for either a single driver to sleep during a part of his journey or no driver at all.
what I would see as an easy starting point would be for auto vehicles to drop their trailers at service stations ..then local drivers act like yard shunters ..just picking up and dropping off at local businesses .

antrobus

globalds wrote: »

...But just like those paying for the first mobiles were companies buying them for key staff.
Then the first users of this tech will be businesses.

Somewhere ..

Australia. RTZ have robot trucks.

globalds

http://www.reuters.com/article/2015/02/14/apple-autos-idUSL5N0VO0L920150214

Technology giant Apple is looking beyond mobile devices to learn how to make a self-driving electric car, and is talking to experts at carmakers and automotive suppliers, a senior auto industry source familiar with the discussions said on Saturday.

Generali

antrobus wrote: »

Australia. RTZ have robot trucks.

They have had for some time. The reason is that it costs a fortune to employ people in Australian mines. Take the Pilbara in WA: nobody lives in the region apart from the very occasional nomadic Aboriginal perhaps so everyone is a Fly In, Fly Out employee. Costs for food, housing, water even are huge as it is the middle of the desert.

cells

antrobus wrote: »

Apparently not.

The cost of refining is about Which is the equivalent of less than $4 a barrel.

http://www.bbc.co.uk/news/business-15462923

The cost to refine petrol is low according to the eia but the cost of diesel refining is much higher.

the eia puts diesel refining at $22 per barrel equivalent with petrol at below $2 not sure why its such a huge difference maybe it's the cost of removing sulphur from diesel?

Anyway the eia also shows that crude is only about 50% of the cost taxes 15% and other things 35% and that was at $70 oil. At $50 the final cost not including the low taxes in the states is about half crude half other

cells

Generali wrote: »

That kinda implies down to $54,000 inis big and costly5-6 years and down to under $20,000 in a decade if the cost goes in a straight line.

If we move to a model of cars on demand, perhaps that cost is being shared across 3 or 4 people. Over a decade lifetime of a car that's a couple of bucks per person per day.

Most likely the hardware will not be the 3d laser system that is big and costly but will migrate to HD cameras. The 3d laser system might have a problem if there were multiple such cars on the road. How would it know the reflected light is from its laser rather than the car next door etc?

project tango is a 3d dual camera smartphone google is working on. It maps 3d space using two cameras and an accurate accelerometer pretty much how our two eyes and perception of movement work. That is effectively just a smartphone and its future cost is probably not more than $200 that is the type of hardware and cost I would imagine a future robocar will work with


One problem is computational power. Although computers are fantastically fast at the moment and can solve huge math problems lile modelling the weather they do so over days or weeks or hours.

a car needs to probably calculate and decide what to do every 1/100th of a second. That itself is probably trivial. The part that is perhaps much more dificult is a program understanding the imagine fed to it in just 1/100th of a second.


Somehow the human brian is great at finding waldo. But a computer is terrible at it and it needs to get good and do it in 1/100th of a second.

cells

antrobus wrote: »

Yes, like I said "if you can shift to mass production you can really get the cost down".

Basically, as with any new tech, you need customers who are willing to pay what (in retrospect) is an extortionate price for the product, in order to kick-start the market. (I mean, three thousand pounds for a Motorola 8000X and you can't even do text messaging; you're kidding right!:)) But yet, there were people out there willing to pay that kind of money just for the convenience of being able to make a phonecall from practically anywhere.

I suppose you could say that with a robot car; you don't need a driving license to drive a car, you don't need to be sober to drive a car, you don't need to be physically capable of driving a car to drive a car, and it doesn't matter about your NCD, or how many accidents you've had in the past five years, or your age, sex, or anything; insurance is still only £50 a year.

Maybe there will be enough ladies-who-lunch who fancy the idea of an instant on-demand taxi parked on the driveway, alcoholics, senile half-blind OAPs, teenage boys, and just plain bad drivers, who will at least consider paying £20k for the equivalent of a robotised Hyundai i10 as opposed to £10k for a normal one, and get things off the ground.

The taxi companies will take it off the ground. Self drive tech at £50k for you or me might be far far too costly but for a taxi company its probably only 1-2 years worth of wages/employer taxes etc for a car they might have for 10 years...

but having said that imo all the tech that will be needed already exists or will exist in mobile phones so the cost cutting and the volume is already there.

eg GPS already exists
HD high frame rate cameras already exist
3G/4G mobile Internet already exists
wifi already exists
high accuracy accelerometers already exist
Laser distance measurement exists
Very fast cpus and gpus exist
digital maps and mappig exists

All of those currently cost just dollars and even all together its a price tag in the region of $100


The only lacking piece is the software and perhaps the computational power for the software to understand the imagines fed to it. And ince the siftware exists to copy it costs close to nil

antrobus

cells wrote: »

The taxi companies will take it off the ground. ...

Possibly. But in the long run the Zombie Robot Driving Apocalypse will destroy the taxi market. When every car is driverless, every car is potentially a taxi.