The Self-Driving Vehicles Thread

[Horse]

I'm struggling to see how that works. There's a more or less fixed market for cars, so the longer a car lasts the fewer NEW vehicles are sold.

Whether or not it's *the* answer, I don't know., but there's a market in developing countries for first World cast-off vehicles. Associated with this is lower crash worthiness and emissions.

NB IIRC in India there are cars manufactured and on sale which would achieve '0' of NCAP testing!

In parallel, you might have noticed that NZ has few new cars, but last of Japanese used cars imported. This is a concern as some people fear they will end up with leftover ic engined cars and left behind in the transition to EVs.

[Mr. Dazzle]

Consider also how cars are sold now. Most are sold on 2-3 year PCP deals, then sold again on a PCP deal as an "approved used" car and then AGAIN and so on. So it is in the interest of manufacturers to make cars with a decent life time. Remember also that the value of a PCP deal is based on the assumed residual value of the car, so if you make a car which holds its value you can offer better PCP deals and sell more of them.

I'm struggling to see how that works. There's a more or less fixed market for cars, so the longer a car lasts the fewer NEW vehicles are sold. IIRC that was the problem that Volvo ran into in the 70s and 80s - cars that easily last two decades! They revised the vehicle lifetime downwards. Renault, Citroen and Peugeot took this so far that some of their early 2000s vehicles never made it past the first UK MOT.

Only if you assume people just buy a new car when they NEED one.

That goes as much for people buying a brand new ca or a car that's just new to them.

There's also not a fixed market for cars, otherwise there would be the same number of cars on the road as there's always been wouldn't there?

[Mr. Dazzle]

Should also add that planned obsolescence is deffo a thing, but that's not the same as planned failure.

[Mr. Dazzle]

Tech that doesn’t work

https://www.thisismoney.co.uk/money/car ... ction.html

People often point out failures like this. For balance you should probably also include stories like this

https://www.bbc.co.uk/news/uk-england-m ... r-55666013

Technology and People both have the foibles of course, we can all point and laugh at examples of both doing dumb shit.

One key difference though is that you generally only have to fix the tech once! People (or indeed any one person) will make the same mistake over and over and over again. In fact people will go one better and deliberately make mistakes on the grounds of "you can't tell me what to do". They may even be inspired to go out of their way to make mistakes by watching someone else. Tech seldom does that.

[Horse]

Tech that doesn’t work

https://www.thisismoney.co.uk/money/car ... ction.html

People often point out failures like this. For balance you should probably also include stories like this

https://www.bbc.co.uk/news/uk-england-m ... r-55666013

Or low-tech that you would have was easy to get right?

www.bbc.co.uk/news/amp/business-43987134

The first reported incident of a Zafira catching fire was noted by Vauxhall in 2009, but it took until August 2015 for the carmaker to launch an investigation, after recording 161 reports of fires.

In some cases, the fires were serious enough to destroy the entire vehicle, and victims were left with mere seconds to escape. To date, no one has been seriously injured in one of the vehicle fires,

In November 2015, Vauxhall issued a recall for 200,000 Zafira B cars , and then a second recall for 234,000 cars of the same model in May 2016.

The carmaker said the fires were caused by "unauthorised repair" of the thermal fuse in the vehicle's blower resistor.

Small point: was it the "unauthorised repair" that caused the fires? Well, 161 reported fires, so how many unreported? Out of how many repairs - did they all result in fires? The initial problem was a poor design thermal fuse which, you might hope, had been recognised as safety critical. Then they had not one, but two, recalls.

[The Spin Doctor]

Only if you assume people just buy a new car when they NEED one.

What I am assuming is that there are only so many cars anyone can drive!

That goes as much for people buying a brand new ca or a car that's just new to them.

There's also not a fixed market for cars, otherwise there would be the same number of cars on the road as there's always been wouldn't there?

Errr... the population has grown.

But actually, although there are more multi-car households than in 2020, the number of cars on the road hasn't grown as much as you might think.

[Horse]

the number of cars on the road

That's an interesting point. Pick an airfield, any old disused airfield.

20210115_210500.jpg (267.79 KiB) Viewed 27858 times

20210115_210433.jpg (344.72 KiB) Viewed 27858 times

20210115_210531.jpg (192.31 KiB) Viewed 27858 times

[Dodgy69]

Only if you assume people just buy a new car when they NEED one.

What I am assuming is that there are only so many cars anyone can drive!

That goes as much for people buying a brand new ca or a car that's just new to them.

There's also not a fixed market for cars, otherwise there would be the same number of cars on the road as there's always been wouldn't there?

Errr... the population has grown.

But actually, although there are more multi-car households than in 2020, the number of cars on the road hasn't grown as much as you might think.

Our roads must be shrinking then.

[wheelnut]

People often point out failures like this. For balance you should probably also include stories like this

https://www.bbc.co.uk/news/uk-england-m ... r-55666013

Technology and People both have the foibles of course, we can all point and laugh at examples of both doing dumb shit.

One key difference though is that you generally only have to fix the tech once! People (or indeed any one person) will make the same mistake over and over and over again. In fact people will go one better and deliberately make mistakes on the grounds of "you can't tell me what to do". They may even be inspired to go out of their way to make mistakes by watching someone else. Tech seldom does that.

And yet we still haven’t yet managed to make a car that can drive itself, despite these flawed humans generally managing to do it every day.

Nor does it look like we’ll manage it anytime soon.

[Cousin Jack]

And yet we still haven’t yet managed to make a car that can drive itself, despite these flawed humans generally managing to do it every day.

Nor does it look like we’ll manage it anytime soon.

It's technology innit. It fails - people don't fail, it's the other bloke's mistake.

[Mr. Dazzle]

And yet we still haven’t yet managed to make a car that can drive itself, despite these flawed humans generally managing to do it every day.

Nor does it look like we’ll manage it anytime soon.

I don't think that's quite true. In the situations where we allow cars to drive themselves they're already safer than humans, in pure statistical terms.

[Horse]

And yet we still haven’t yet managed to make a car that can drive itself, despite these flawed humans generally managing to do it every day.

Nor does it look like we’ll manage it anytime soon.

People have been driving powered vehicles since 1769, with big increases in the numbers of 'ordinary' drivers gaining experience during WWII and a massive rise in ownership from the 1960s onwards.

Although technology is accelerating at a phenomenal rate (iPhones with Lidar, for example, rather than a £10k gizmo on the roof of an AV), but driving has also become incredibly complicated during the last 50 years (traffic density, road complication, etc.).

Although I don't see the full L5 autonomous vehicle (as someone said, a truly autonomous vehicle would say "Go to the office? Nah, I'm off to the seaside for the day") happening in my lifetime, but we already have had incredible advances in what can be done. That said, it's an incredibly complicated thing to try and achieve.

Don't limit your perception of this solely to road vehicles. I'll sort out some examples of other stuff tomorrow.

[Dodgy69]

Can we have a pole.?

Would you travel in a self driving car on our current road network.?

I don't know the numbers, but I reckon accidents are still pretty rare when considering vehicle numbers on our roads. They are in my little world anyway.

[demographic]

I'm not convinced that self driving cars will be the first real self driving vehicles.
Haulage can make huge gains by eliminating the driver or having several vehicles semi autonomous in a close train style with less drivers.
Quarry trucks, steelyard shunt lorries and more trains.
Lorry drivers may well be hit hard by this, the ones doing unusual routes and handling at each end will survive but you can bet yer arse that Stobarts and others will be looking hard at this.

Just taking someone to work when a lot of people actually enjoy driving? Dunno, I wouldn't pay extra for a car that did that.

[Mr. Dazzle]

I agree, I think commercial vehicles will be the first mass adopters. If nothing else, not having to pay a guy is quite attractive to firms.

Its also much easier to set up depots for autonomous vehicles, so 'warehouse to warehouse' type stuff, where you only drive on motorways and similar, will probably come along first.

[Horse]

I'm not convinced that self driving cars will be the first real self driving vehicles.

Quarry trucks, steelyard shunt lorries and more trains.
Lorry drivers may well be hit hard by this, the ones doing unusual routes and handling at each end will survive but you can bet yer arse that Stobarts and others will be looking hard at this.

You're very right. There are already many non-roads applications of AVs. The military want them (IEDs etc), parcel delivery (Dazzle's MK bathtubs on wheels), mining/quarrying already has them, trains, warehouses, agriculture. See vids later in post.

The on-road stuff is higher risk because it's a totally uncontrolled environment. But low-speed shuttle bus applications may not be too far off. I have been a passenger in an automated vehicle, a couple of years ago. At that time, it wasn't good enough to go on-road.

Couple of videos.

The SMLL is long, but worth watching entirely as it shows how a number of strands are being woven together including: digital twin environment for testing AV control software against multiple variations of known crashes using far more detailed information from in-depth analysis of specific crashes.

The other two videos show how things are progressing. That said, the UK's implementation and understanding of 'safety drivers' is far different to that in the USA. The Uber crash [probably] wouldn't have happened here.

SMLL


Tesla autopilot


AV no safety driver


Other applications:

https://techcrunch.com/2021/01/05/oxbot ... lications/?

However, and perhaps counterintuitively, automation is being led by agriculture, in some areas a decade ahead.

Harper Adams University started 3 years ago with their hands free hectare project, no-one entered the area during the year while they went through the entire process to plant, grow, then harvest a crop. The vehicles used were auonomous from leaving the 'barn'. Improved in 2019, then several large fields in 2020:

[Dodgy69]

I'd be a happier motorist if say, HGVs had their own lane with barriers to keep other motorists safe and reasured. Maybe to run alongside our motorways, but then there's still the off motorway bit to sort out.

For the sake of a drivers poor wage, I'm unconvinced its really worth the hassle. I'm not against new tech, but it must be progressive, reliable and cost effective to be a goer.

[Horse]

For the sake of a drivers poor wage, I'm unconvinced its really worth the hassle. I'm not against new tech, but it must be progressive, reliable and cost effective to be a goer.

See my post, 'roads' are just a small part of it.

And, don't forget, many of the elements are already in new cars: cameras, sensors, radar, etc., to provide systems such as adaptive cruise control, collision warning, auto braking, lane keep & warning, self-park, stability control, probably more.

[Dodgy69]

The electrickory is a great safety backup, but to have your life depend on it is different. I honestly can't see them sharing our public roads in real world, any time soon.

[Horse]

The electrickory is a great safety backup, but to have your life depend on it is different. I honestly can't see them sharing our public roads in real world, any time soon.

We bought a used Ibiza in 2001 that had 'fly by wire' throttle (I know this because a connector failed and it went hay[noflyby]wire).

If you have a relatively modern car, you may be trusting electronics more than you realise.

But I know what you mean.