Fully automated vehicles - 'not in our lifetime'?

GreatApe

Arklight wrote: »

Considering the "self driving" Uber car that ran over that poor homeless woman in Florida apparently did so because it failed to recognise her as human, I think we're a looooong way from autonomous vehicles being able to make complicated philosophical judgements about the existential value of different types of life.

In reality this AI technology knows roughly where it is through GPS, kind of hopes what it can determine of the road is the same as the maps it has, and sees blobs meandering around which could be anything.

I think if we can accept that this is acceptable driving and that AI drivers have no hope of passing the hazard perception part of the UK driving test (there's a child holding a ball and a distracted mum!), then the tech could be ready for prime time within a decade.

Otherwise, not in the lifetime of many of the users of this board.

Far too pessimistic the prize is so large ($10 trillion +) for the company that cracks it that billions and billions of dollars and man hours will be spent on it

Its also a software problem all the hardware is already in place.
A self drive car has senses far above human senses. More eyes more ears at higher resolution and spectrum radar lidar ultrasound.

With AI we don't even need humans to program them mass data and AI will program them.

And these don't have to be perfect or flawless they just have to be better than humans who are pretty terrible drivers considering the deaths and injuries

Herzlos

GreatApe wrote: »

The majority of people don't live on top of a station and want to get to on top of another station.

If you go from say Watford to Birmingham say the south east of Birmingham its what a 100 mile drive that's a 1.5 hour drive.

To do the self drive taxi plus trains you need to first drive into central London at 30mph going the opposite way for 15 miles and half an hours time. You then need to leave the taxi and walk to the station and get a ticket and wait for the train let's say that's 12 mins. So your already down 42 minutes when you get on the train its another 1 hour 25 minutes to Birmingham new Street. Your journey to new street took you over two hours then maybe another 15 mins to SE Birmingham making 2 hours 20 mins

So in this example its 1h 30mins for the taxi or 2h 20mins for the train

The taxi costs £9 in fuel and about £15 in capital costs = about £25 for a ticket assuming you don't share the trip with anyone. Share with 4 persons and its £7 a ticket

Compared to £25 a ticket for the cheapest off peak train ticket today (plus a self drive taxi to and from the station)


So the taxi would be an hour quicker and less then half the price

Sure you can easily pick edge cases where people are quicker driving than getting a train, where they are likely already driving. You're also assuming that the best train route is to drive half an hour in the wrong direction at each end.


You could easily get a self driving taxi out to a connecting train station, timed to get you on that train in plenty of time, then get off somewhere south of Birmingham and straight into a self driving car at the other end.


Lots of people visiting other cities are doing so within walking distance of the main train stations. People currently getting a train from their home in the suburbs to an office in the city centre aren't going to suddenly start taking self driving cars - the congestion would be incredible.



You're also assuming that we couldn't have more train stations in the future.


Japan has train stations everywhere, and it's not uncommon to have a journey with 4+ train changes. It's actually really easy when the trains are so regular and consistent - you can time it to within about 30 seconds, and if you miss it there will be another one in a few minutes.

Herzlos

GreatApe wrote: »

Its also a software problem all the hardware is already in place.

Not really. The hardware for the test mules is in place, but it's not ready for the prime time yet, unless you're happy for most of the boot space to be computer equipment and battiries.


Another 10-20 years and you'll barely notice the processing equipment.

Arklight

GreatApe wrote: »

Far too pessimistic the prize is so large ($10 trillion +) for the company that cracks it that billions and billions of dollars and man hours will be spent on it

Its also a software problem all the hardware is already in place.
A self drive car has senses far above human senses. More eyes more ears at higher resolution and spectrum radar lidar ultrasound.

With AI we don't even need humans to program them mass data and AI will program them.

And these don't have to be perfect or flawless they just have to be better than humans who are pretty terrible drivers considering the deaths and injuries

I actually agree with you that on balance the mediocre driving that AI is capable of would save more lives than the wildly inter and intra differential standards of human drivers.

In practise it cannot assess real world conditions for the same reason there is no robot on earth that can walk into your house, find your kitchen sink, find your washing up liquid and sponge, wash up a coffee cup for you, and leave.

AI sort of works with driving because roads sort of remove most of these types of variables.

The Uber that killed the pedestrian in Florida had lidar yaddah yaddah. It detected her on the road, had no idea what she was because she was pushing a bike so just looked like a funny shape, and drove right into her. If Uber now reprogram their vehicles to jam on the anchors as soon as they don't recognise something on the pavement or road none of them will move an inch.

There is no AI either existing now or in the pipeline that can look at a room full of people or a busy street and understand what it is seeing. You can't drive safely unless you can look at road, the condition and orientation of the other vehicles and drivers, and all the pedestrians on the sidewalk and understand what you are looking at. A decent driver on a good day can, and is legally obliged to. End of.

antrobus

GreatApe wrote: »

...

It will also kill point to point systems like trains.

Leading to a significant increase in road congestion.

Told you so.

GreatApe

Herzlos wrote: »

Sure you can easily pick edge cases where people are quicker driving than getting a train, where they are likely already driving. You're also assuming that the best train route is to drive half an hour in the wrong direction at each end.

You could easily get a self driving taxi out to a connecting train station, timed to get you on that train in plenty of time, then get off somewhere south of Birmingham and straight into a self driving car at the other end.

Lots of people visiting other cities are doing so within walking distance of the main train stations. People currently getting a train from their home in the suburbs to an office in the city centre aren't going to suddenly start taking self driving cars - the congestion would be incredible.

You're also assuming that we couldn't have more train stations in the future.

Japan has train stations everywhere, and it's not uncommon to have a journey with 4+ train changes. It's actually really easy when the trains are so regular and consistent - you can time it to within about 30 seconds, and if you miss it there will be another one in a few minutes.

It is not just the time but also the cost and convenience I can see self drive taxis costing less than half of what it would cost to take a train on economy ticket and a non shared self drive taxi in a nice vehicle will cost a lot less than a first class train ticket. As fewer people use the trains they will also cost more per passenger mile while robo taxes can expand or contract with demand. And it isnt just trains it is also some domestic flights too

The robo taxis also include highly taxed fuel so are net contributors to taxes while most trains even today need to be subsidized.

You could have a future where the old train lines are converted to roads I think that is most likely. The stations can become housing or offices the lines could become new roads linking cities.

Malthusian

GreatApe wrote: »

With AI we don't even need humans to program them mass data and AI will program them.

After all the rubbish spoken about AI on these boards, this is the first case of someone suggesting something that could actually be feasibly be done with what is currently marketed as AI. I can imagine an AlphaZero type program being given a game to play where the gameboard is a simulation of a road network with lots of blobs moving about, and the objective is to get from any point A to B a billion times without hitting any pedestrians. DriverZero is never told which blobs are pedestrians until it hits one, in the same way as AlphaZero was never told that 1. e4 is a good idea and 1. f3 2. g4 is not. Its purpose is to run through algorithms until it finds one that allows it to correctly identify pedestrians and navigate from A to B a million times at a reasonable speed without hitting any.

The problem of course is that we have only just managed to invent an "AI" that can teach itself to play chess. Chess has 64 squares and 32 pieces which move in 6 different ways. A 3D street scene with humans milling about in all directors is an unfathomably more complex game. Whether it's feasible to have an AI which is capable of teaching itself to play "don't hit the pedestrian" in the same way that AlphaZero taught itself chess, I honestly don't know. It's quite possible that humans will figure it out before we have a powerful enough computer.

Herzlos

You're assuming that robo taxis will be cheaper than human drivers, not factoring in how much the robo taxi will cost, or that fuel is only going to go up in price (or be electric and have issues there).


Trains are still going to be better for longer distances (over an hour), or anything that has a large number of passengers.


Trains do cost too much though; but that's a separate problem.


As for turning train lines into roads - where will the freight go? I doubt the tracks would make good roads in a lot of cases because they can't be widened enough.


Trains aren't going anywhere. They may look totally different in the future but they still provide the best transport options for hub-hub or satellite-hub transit.

GreatApe

antrobus wrote: »

Leading to a significant increase in road congestion.

Told you so.

I doubt it because I suspect we will go from a median car carrying just 1 person to the median robo taxi carrying 3 persons and there will likely be peak and off peak pricing. Even the peak pricing will be cheaper than owning your own car but the off peak could be even half of the peak pricing. I suspect the intercity train lines will mostly be abandoned and perhaps converted to roads if the capacity is needed but I dont think the capacity will be needed

Another factor is the cost of buying and insuring a car when you are young is very high.
What does it cost these days maybe £1,000 in lessons, £2,000 first year insurance, £2,000 old crappy car, £1,000 fuel and maintenance etc? So £6,000 upfront to do maybe only 10,000 miles so 60p a mile but all of it upfront. Why do that if a robo taxi costs 20p a mile and is a nice shiny nearly new car? Fewer kids will opt to learn and pay a huge sum upfront and just opt to continue using the robo taxis they have been using since age 10

GreatApe

Herzlos wrote: »

You're assuming that robo taxis will be cheaper than human drivers, not factoring in how much the robo taxi will cost, or that fuel is only going to go up in price (or be electric and have issues there).

Trains need fuel too the London underground iirc uses something like 100g CO2 per passenger KM that is close to a modern car with just 1 passenger in it. That means a taxi with 4 passengers uses 75% less energy than a train system. The trains themselves might be efficient but the whole system (eg to power that stations and everything else) added together isnt very efficient

Trains are still going to be better for longer distances (over an hour), or anything that has a large number of passengers.

Not really because a robo taxi can also scale in size you can have a large robo taxi eg a robo
8 person people carrier or even a robo bus with 200 passengers more importantly the robo vehicle system can use smart data to operate at higher capacity. While a train with a capacity of 1000 might have only 50 passengers at off peak times

Trains do cost too much though; but that's a separate problem.

Trains dont cost much, the train line infrastructure costs too much and there isnt much that can be done to improve that. HS2 is going to be a super silly amount of money. The robo taxis will cost much much less and can expand contract or even be deployed elsewhere.

As for turning train lines into roads - where will the freight go? I doubt the tracks would make good roads in a lot of cases because they can't be widened enough.

Freight will go onto self drive electric HGVs

Trains aren't going anywhere. They may look totally different in the future but they still provide the best transport options for hub-hub or satellite-hub transit.

Sure I dont expect them to die overnight
But there could be a very sharp drop in passenger numbers which leaves the government with two options. Increase ticket prices further or subsidies the trains even more either of which is going to be pretty. Why sub trains billions more when you can put it into heathcare? And if you let the customers pick up the additional per passenger cost of fewer customers then more will leave the trains and go onto the robo taxis/buses