Green, ethical, energy issues in the news

joefizz

Martyn1981 wrote: »

No nice way to put it, Tesla scares the ..... out of the old firm

Spend any time in one and its obvious why.
Tesla is an electric car.
The rest are cars with an electric drivetrain. Everyone else is now playing catchup.

GreatApe

joefizz wrote: »

Spend any time in one and its obvious why.
Tesla is an electric car.
The rest are cars with an electric drivetrain. Everyone else is now playing catchup.

Everyone is trying to get to self drive first that is the prize

All the large automakers will have compelling BEVs
There is no magic to batteries or electric motors
Both are much simpler than combustion engines
So all the big auto companies will get there
It doesn't matter if they are 3-5 years late
At this stage BEVs are such a small fraction of market share they ain't missing out on much

The prize is self drive software
There it very very much matters if you're 18 months after the first guy you've already lost
Not so with BEVs the other companies it doesn't matter much if they are five years late to BEVs

1961Nick

GreatApe wrote: »

Everyone is trying to get to self drive first that is the prize

All the large automakers will have compelling BEVs
There is no magic to batteries or electric motors
Both are much simpler than combustion engines
So all the big auto companies will get there
It doesn't matter if they are 3-5 years late
At this stage BEVs are such a small fraction of market share they ain't missing out on much

The prize is self drive software
There it very very much matters if you're 18 months after the first guy you've already lost
Not so with BEVs the other companies it doesn't matter much if they are five years late to BEVs

It could still take a lot of years to get to fully autonomous capability. The problem is that 99.999% isn’t good enough ... it has to be 100% for obvious reasons.

JKenH

1961Nick wrote: »

It could still take a lot of years to get to fully autonomous capability. The problem is that 99.999% isn’t good enough ... it has to be 100% for obvious reasons.

A lot of years before it will be able to cope with single track Lincolnshire back roads with ill defined edges. Can autonomous BEVs overtake tractors on country roads?

GreatApe

1961Nick wrote: »

It could still take a lot of years to get to fully autonomous capability. The problem is that 99.999% isn’t good enough ... it has to be 100% for obvious reasons.

Doesn't have to be 100% it has to be as good as a human if not better but humans are nowhere near 100% as obvious by the number of crashes

But yes it's a very difficult problem
But it will be solved at some stage there is just too much money to be made from this software and a lot of talent going into trying to solve it

I think it will happen in the 2020s it's anyone's guess when exactly but I'd guess the midpoint of the decade

GreatApe

JKenH wrote: »

A lot of years before it will be able to cope with single track Lincolnshire back roads with ill defined edges. Can autonomous BEVs overtake tractors on country roads?

Might start as mini delivery trucks first to prove the concept

Imagine a ford transit but much smaller instead of a 1.5 ton self drive car we begin with a 100kg self drive delivery truck. Imagine something the size of a dodgems. If they hit a car absolutely no problem if they hit a human then very unlikely to do major damage

Deploy it in say one of the London boroughs with a 20mph speed limit
Let this drone work for the food delivery companies and the small parcel deliveries

Such a small vehicle would pose no danger to other vehicles
Using them in a borough with a 20mph speed limit would make them very safe

They will have accidents but they would get better and better and when their accident rate per million miles is below humans you can scale them to larger sizes

You can also have a remote human able to take over when/if necessary

mmmmikey

Lot's of discussion and excitement about autonomous cars, but I keep asking myself what problem will they actually solve? How will life be better with autonomous cars - is this something we actually need or is it something we are pursuing simply because we can?



Wouldn't it be better to put our efforts and energy into improving schools so children can walk to the local school avoiding the need for the school run? Or sorting out the housing market and public transport so the daily commute becomes a sociable activity on the bus rather than a slog in the car? Why not change our shopping habits and have our groceries delivered and pick up the odd pint of milk in between times from the corner/village shop avoiding the need for the supermarket trip - and so on.


The car has pretty much become a necessity of modern living and of course will continue to be for some people, but in an awful lot of cases there's no inherent reason for this. We've adapted our lifestyles over the last 30 or 40 years to make good use of the car but there's no inherent reason it needs to be that way.



There are aspects of the motor industry that seem to me somehow reminiscent of the predictions around the paperless office - something that has been much talked about, is technically feasible but for a whole host of reasons has never really happened.....

joefizz

GreatApe wrote: »

Doesn't have to be 100% it has to be as good as a human if not better but humans are nowhere near 100% as obvious by the number of crashes

But yes it's a very difficult problem
But it will be solved at some stage there is just too much money to be made from this software and a lot of talent going into trying to solve it

I think it will happen in the 2020s it's anyone's guess when exactly but I'd guess the midpoint of the decade

Its not the software, its the data. Seriously, go sit in a Tesla and see what I mean.
The leaders are Tesla, Uber and Lyft, google maps traffic etc, the rest are so far behind (except surprisingly Volvo) because they dont have data of their own. The best self driving software is useless without the data.
Thats why I changed my mind about Tesla, that and every single place I stopped had a Tesla fast charger in the US. The cash burn also funded the infrastructure, which again the others are relying on 3rd parties to provide.



In the US things are different as I mentioned on this forum before, wide open roads, cities with grid systems, people who pay attention to crosswalk signals etc...
Get on a backroad here and it cant differentiate between edge lines (there arent any) and overgrown hedges or know to drive in the middle of road during sileage/muck spreading/harvest seasons. On the motorways, great, dual carriageways, great.
Watch the screen discriminate between small and large cars passing but would have no idea how to deal with a tractor with lots of death dealing implements hanging off the back.
When the car decides the road is 40mph when its actually 60 and vice versa because its maps or data have never been down that road before. There was an app a few years ago that paid people to go out and drive the backroads, cant remember what it was but I think they disappeared but they were ahead of the game.

Every so often it has a bit of a wobble and asks you to take the wheel or occasionally asks you to press the brakes, jiggle the steering wheel etc.
Driver assist with hands on the wheel isnt that far away, autopilot in this country, maybe never.
Theres always new data, like in the US when you drive on non interstates you have 'in x hundred yards or half a mile it will be 40' and the Tesla thinks thats 40 and hits the anchors. How do you override? The cameras seeing a 40 or the satnav/google traffic saying 40. The cameras (obviously) in case its a temporary speed limit, or is it obvious its not picking something up reflected or a sign painted on the back of a HGV or bus?
Software is garbage in, garbage out and design decisions/assumptions will only get you so far (ask Boeing).

Dont forget every new line painting, speed bump, speed restriction etc etc is probably because one or a number of people has been wiped out. As is every coding error. Ive mentioned here before meeting the people who wrote autopilot software who wouldnt fly on those planes...
...if I were a betting man Id say fully autonomous driving will never happen in the UK/Ireland as a whole but will be allowed on restricted areas like motorways/dual carriageways, might even be legislation to automatically turn it off as you take an exit. Certainly can see it useful in parking mode in car parks where it acts as valet parking provided no humans walking around.



Bringing it back on topic, watch the news items to see who gets to pick over the pieces of Uber/Lyft etc rather than any car company, its the data that is driving all this (excuse the pun). The rest will have to license the software...

I'll reiterate, if you get the chance to go out in an UK spec Tesla and can get past fart mode on the screen it really is a glimpse of the future, then go sit in an i3 or whatever and see the difference. Their self driving semi should make a killing in the US.

joefizz

mmmmikey wrote: »

The car has pretty much become a necessity of modern living and of course will continue to be for some people, but in an awful lot of cases there's no inherent reason for this. We've adapted our lifestyles over the last 30 or 40 years to make good use of the car but there's no inherent reason it needs to be that way.



There are aspects of the motor industry that seem to me somehow reminiscent of the predictions around the paperless office - something that has been much talked about, is technically feasible but for a whole host of reasons has never really happened.....

This.
The move to EVs and (as with the BEV thread discussion) reduction in personal ownership through credit withdrawal etc etc and every government doing more QE only this time on infrastructure (public transport).
A drop in oil prices due to recession then massive increase in prices afterwards will also kill off a lot of it, taking us back to walking around and going by public transport.

Or simply just legislation or green deals will do the job.
The running to the shop for a pint of milk can be done by walking or electric moped/bike (if they ever get round to legalising them!) I saw a meme the other day saying the UK used to have the worlds largest fleet of electric delivery vehicles (milk floats!) and what happened to them?

GreatApe

mmmmikey wrote: »

Lot's of discussion and excitement about autonomous cars, but I keep asking myself what problem will they actually solve? How will life be better with autonomous cars - is this something we actually need or is it something we are pursuing simply because we can?

Wouldn't it be better to put our efforts and energy into improving schools so children can walk to the local school avoiding the need for the school run? Or sorting out the housing market and public transport so the daily commute becomes a sociable activity on the bus rather than a slog in the car? Why not change our shopping habits and have our groceries delivered and pick up the odd pint of milk in between times from the corner/village shop avoiding the need for the supermarket trip - and so on.

The car has pretty much become a necessity of modern living and of course will continue to be for some people, but in an awful lot of cases there's no inherent reason for this. We've adapted our lifestyles over the last 30 or 40 years to make good use of the car but there's no inherent reason it needs to be that way.

There are aspects of the motor industry that seem to me somehow reminiscent of the predictions around the paperless office - something that has been much talked about, is technically feasible but for a whole host of reasons has never really happened.....

What self drive technology allows is to increase the utilisation of cars

So instead of the UK having 35 million cars used 5% of the time we might go to 12 million cars used 15% of the time. This is a massive increase in productivity. It means we need to manufacture only 1/3rd as many cars per year.

That alone saves the UK a massive sum of money. Going from 2.5 million cars at say £25k to 0.8 million cars at £30k each is a near £40 billion per year saving for UK consumers or close to £1,500 per household.

It's in fact more than this. For instant how much do kids spend learning to drive? Let's say £1k each and 800,000 people learn led year this is another £0.8 billion per year saving.

And ideally these will be safer than humans. If the average insurance is £500 and say half that goes to fix accidents and accidents fall by half that's about a £5 billion saving

Of you keep adding up the savings it comes to somewhere around £60-80 billion

So why do we need self drive cars?
Because it allows us to divert £60-80 billion in resources to something else
That's like hiring a million more doctors

It's a huge huge prize the biggest near term prize you can imagine