Autonomous Vehicles

[Deleted User 234324]

I think 10 years is a perfectly reasonable time frame for mainstream autonomous vehicles, when you consider that the likes of the Tesla model S/3/X already have the required hardware for full autonomy and can already navigate well through a town (in the US), it's more a matter of waiting on the software to mature enough. Self driving vehicles are one of those technologies that when they catch on, will have a sudden surge, with the most likely scenario being logistics and transport being the first for mass adoption. I would not be surprised at all if by the end of the next decade, there will be very few if any commercial vehicles on the road that are not autonomous.

It will be one of those technologies that will just suddenly take off, first it will be the distribution trucks that go from a depot to a shop or factory to depot etc, the moment an autonomous vehicle becomes available that tesco, asda etc can use for home delivery, they will be all over it. The savings from staff, insurance costs etc will be colossal.

 

[Amp34]

So Budweiser has done one single trip on a pre planned route. That still needed the driver to do the main bulk of the complicated driving. You are really clutching at straws. In a nutshell that is a development of cruise control. Your original statement was we were going to have automation as soon as 5-10 years in the mainstream which is what I was arguing. We will have automation in the future that is definitely true but not as quick as you are saying.

Yes, the aim was a proof of concept. To provide that a lorry can safely drive several hundred km on a main road without any human intervention. This isn't cruise control at all. It's level four automation (go look it up).

Lets lay out the evidence.

• A proof of concept to show tech is available to drive a lorry from Point A - Point B on a main road has been successfully completed.
• Smaller vehicles are at the stage where they can safely navigate towns and cities, mixed in with regular human traffic, while someone is sat in the back seat, unable to take immediate control.
• It's likely next year some of those vehicles are going to be level 5 autonomous and not even have an employee in the back seat (completely autonomous taxis picking up and dropping off paying customers).
• Companies, including the likes of Tesla, Waymo (Google), Otto (Uber) and the traditional trucking manufacturers are investing hundreds of millions into automation of HGVs, alongside the traditional car manufacturers.
• The UK government/US government and various European governments are currently working on legislation for commercial and personal use of autonomous transportation on the regular road network.
• Removal of a human from the equation will significantly reduce operating costs in haulage, meaning operators are/will be very interested in the cost savings when it's available.

And to clarify. I didn't say anything about "mainstream". I said, and was very clear about it, that I think in 5 years there will be companies testing automated lorries (such as the one that delivered the Budweiser) without any human aboard on open roads, and in 10 years there will be a commercial product that can be fully autonomous (i.e. no human in the cab) hauling along certain specific routes.

No it's not going to be "mainstream", there are unlikely to be hundreds/thousands of those vehicles, and they are unlikely to be in towns or cities. Rather, there may well be a few large goods vehicles that travel autonomously on open roads under a very specific set of circumstances. The rest of it probably is 30 years away, where quite possibly most haulage will be automated, along with taxis and most personal driving.

It's certainly not pie in the sky considering the leaps and bounds autonomous tech has moved forward over the last 5 years, as well as the huge amount of money being spent on it by a significant number of companies and the cost savings it will eventually bring to the haulage industry.

How many large trucking companies would bite the arm off the company offering a vehicle that could halve (or more) their operating costs and allow them to not have to worry about breaks and maximum driving time, allowing them to transport 50% more stuff? That's the sort of savings being banded around with electric and automated haulage.

Again that has nothing to do with what I was saying apart from i mentioned coal. That is tonka trucks moving about a very simple layout in a remote location over a very short distance.

They aren't remote controlled Tonka trucks. Lets remember what you said.

Could a automated lorry deliver coal from some quarry to a refinery in Northern Canada on an empty road in the next ten years? Hell yeah.

That is basically what they are doing already, except most of them are doing it in the quarries. The technology in them would be perfectly capable of moving them along an empty road in northern Canada.

 

[Nasher]

Clarkson recently commented that he had to take back control twice in a 50 mile journey to stop an "auto pilot" system from killing him. It is till a long way off tbh

 

[adam cool dude]

They aren't remote controlled Tonka trucks. Lets remember what you said.

That is basically what they are doing already, except most of them are doing it in the quarries. The technology in them would be perfectly capable of moving them along an empty road in northern Canada.

I didn't say remote control I said remote location. A Tonka truck moving about a very simplistic quarry is different level to a town center. If the technology is there to move along an empty road in northern Canada then why are they not doing it?

 

[adam cool dude]

Clarkson recently commented that he had to take back control twice in a 50 mile journey to stop an "auto pilot" system from killing him. It is till a long way off tbh

Same with that guy who decapitated himself while watching harry potter in his back seat.

 

[VincentHanna]

I believe in all the cases where Tesla's have been in an accident under autonomous driving mode, the fault was with other drivers and not the Tesla?

There's not much the car can do if another road user runs a red light at speed at an intersection or something.

 

[Nasher]

One was a lorry cutting across lanes, but it should have seen it.

 

[Simon]

I believe in all the cases where Tesla's have been in an accident under autonomous driving mode, the fault was with other drivers and not the Tesla?

There's not much the car can do if another road user runs a red light at speed at an intersection or something.

In many countries I’ve been too, drivers are well versed to watch out at red lights for this. Hence people go a few seconds after green.

Can these autopilots do the check sideways to see the risk of other cars as quick?

For me automation will only work when every car is automated and one system controls to whole lot.

 

[Mercenary Keyboard Warrior]

Its already started, in fact it started years ago.
Remember the early ABS, it wasn't superb but it was an improvement for most drivers. There were the self diagnosed super brake people who could beat the system (so they thought) but now would they claim the same?

Take something like city emergency braking, a computer is already overriding the thoughts of the driver (or probably lack of in reality) to stop a car
Active lane assist
Speed sign recognition, lots of the precursor tech needed is already becoming mainstream on mainstream cars

To me the fact is we are making the evolution harder to get to by having to have both driver and driverless on the road at the same time.
I am sure the driverless would function today, but we throw drivers into the mix, I mean daily who doesn't at some point speed, fail to indicate, select a wrong lane, fail to observe braking distance, or all of the above.

One of the most obvious places to start this would be in fact docks for the UK. Most of the containers are dropped onto a rig and that will drive to some warehouse somewhere, no loading mid journey etc to be done. Just program the satnav and allow the computer to go there.

First stage to me is to allow the "driver" to hand over control. Then when most have the technology you change the law to give priority to automated, and make any luddites have to give prioity to computer controlled.

 

[Simon]

Abs is a bad analogy. You don’t have 4 brake pedals.

 

[Nasher]

In many countries I’ve been too, drivers are well versed to watch out at red lights for this. Hence people go a few seconds after green.

Can these autopilots do the check sideways to see the risk of other cars as quick?

For me automation will only work when every car is automated and one system controls to whole lot.

But that opens up a whole new can of worms. It's one big, vulnerable single point of failure. One issue would effect every single car, so it would take just one bug and you could have 1000s of crashes.

It's like a cloud system. Now a hacker only needs to hack one system to get the data of multiple companies or cause total nationwide chaos (like we saw happen to the NHS).

Cars should always be self contained systems with no external network links to vital functions.

 

[jigger]

Its already started, in fact it started years ago.
Remember the early ABS, it wasn't superb but it was an improvement for most drivers. There were the self diagnosed super brake people who could beat the system (so they thought) but now would they claim the same?

Take something like city emergency braking, a computer is already overriding the thoughts of the driver (or probably lack of in reality) to stop a car
Active lane assist
Speed sign recognition, lots of the precursor tech needed is already becoming mainstream on mainstream cars

To me the fact is we are making the evolution harder to get to by having to have both driver and driverless on the road at the same time.
I am sure the driverless would function today, but we throw drivers into the mix, I mean daily who doesn't at some point speed, fail to indicate, select a wrong lane, fail to observe braking distance, or all of the above.

One of the most obvious places to start this would be in fact docks for the UK. Most of the containers are dropped onto a rig and that will drive to some warehouse somewhere, no loading mid journey etc to be done. Just program the satnav and allow the computer to go there.

First stage to me is to allow the "driver" to hand over control. Then when most have the technology you change the law to give priority to automated, and make any luddites have to give prioity to computer controlled.

Just throw life into the mix and the self driving thing is flawed. Fuel on the road, traffic film covering a camera, cattle, high winds. So many variable situations.

 

[Mercenary Keyboard Warrior]

Abs is a bad analogy. You don’t have 4 brake pedals.

You will have to expand on that one, I don't think we ever had 4 brake pedals

The computer is better at this really rather simple thing, that actually has quite a lot of variables when you look at it. I was relating the point to all the "I will always brake better than the guy with ABS, I can cadence brake 4,000,000 times a second" that so many so called expert drivers trotted out when ABS became available. Funny that in the insurance game there was a term, "ABS back end" which was from all these super drivers basically stacking it into the back of cars that had ABS
Nowadays no one thinks they can out brake ABS, all the driver techs will go the same way, initially maybe there will be better drivers than the systems, but the systems will soon take over and exceed the abilities of us terrible humans.

 

[Simon]

ABS can control each brake individually. This is where it’s benefit is. The ability to remove pressure from a locked wheel only.

No human can do that with one pedal

 

[Nasher]

A computer can carry out pre-written instructions very quickly. But it can't think for itself or adapt to changing situations and that is what's needed on the road. True AI is the missing piece but that could still be 100 years away, if it's even possible. Even then it might get banned if Hawking etc. get their way.

 

[Mercenary Keyboard Warrior]

Just throw life into the mix and the self driving thing is flawed. Fuel on the road, traffic film covering a camera, cattle, high winds. So many variable situations.

Take fuel on the road, you seriously think you can detect that and react faster than a computer? You may, and very much emphasis on the may, spot it faster but you won't react to it quicker.

What if its pouring with rain, in the dark when your struggling to see through the oncoming vehicles headlights bouncing off the road, when just beyond your lights someone steps into the road, who is going to "see" and react to that situation better?
You can always find things humans will do better for now, the thing is we are just about maxed out now, computers and systems are evolving and at such a tremendous pace that sooner or later everything we do will be able to be done better by a £1 chip.

Take the city braking I mention above, it will stop collisions compared to a driver, we can do one thing at a time, you can have as many computers as you want within reason, all multi tasking if you want, they do that well., that one chip is just sitting there when in traffic meeting x conditions, it just there constantly "looking" for the need to brake. Your average driver will be humming to the radio, or looking at the hot bird in the car to their side, or trying a sneaky text, or to sum it up, not focussing on the same thing as that chip

 

[jigger]

A computer can carry out pre-written instructions very quickly. But it can't think for itself or adapt to changing situations and that is what's needed on the road. True AI is the missing piece but that could still be 100 years away, if it's even possible.

How do you teach AI to deal with a cow in the road or fallen tree? A blowout thats left the remains of the tyre sat 2 feet into the first lane of a motor way.

 

[Mercenary Keyboard Warrior]

ABS can control each brake individually. This is where it’s benefit is. The ability to remove pressure from a locked wheel only.

No human can do that with one pedal

Oh right, yes modern ones yes, iirc the early systems were 2 channel, ie they braked front or rear as a pair. This is exactly the point in how its evolved.

Edit to add clarification, (perfect example of how the early system which already was better than most drivers has evolved)

4) Two-channel, four sensor ABS
This system, commonly found on passenger cars from the late '80s through the mid 1990s, uses a speed sensor at each wheel, with one control valve each for the front and rear wheels as a pair. If the speed sensor detects lock up at any individual wheel, the control module pulses the valve for both wheels on that end of the car.

Now

1) Four-channel, four-sensor ABS
There is a speed sensor on all four wheels and a separate valve for all four wheels. With this setup, the controller monitors each wheel individually to make sure it is achieving maximum braking force.

 

[jigger]

Take fuel on the road, you seriously think you can detect that and react faster than a computer? You may, and very much emphasis on the may, spot it faster but you won't react to it quicker.

What if its pouring with rain, in the dark when your struggling to see through the oncoming vehicles headlights bouncing off the road, when just beyond your lights someone steps into the road, who is going to "see" and react to that situation better?
You can always find things humans will do better for now, the thing is we are just about maxed out now, computers and systems are evolving and at such a tremendous pace that sooner or later everything we do will be able to be done better by a £1 chip.

Take the city braking I mention above, it will stop collisions compared to a driver, we can do one thing at a time, you can have as many computers as you want within reason, all multi tasking if you want, they do that well., that one chip is just sitting there when in traffic meeting x conditions, it just there constantly "looking" for the need to brake. Your average driver will be humming to the radio, or looking at the hot bird in the car to their side, or trying a sneaky text, or to sum it up, not focussing on the same thing as that chip

Yes I can notice fuel on the road a lot of times. Not always of course. I could probably react better than most computers and probably in ways nobody would program for.

 

[Mercenary Keyboard Warrior]

Yes I can notice fuel on the road a lot of times. Not always of course. I could probably react better than most computers and probably in ways nobody would program for.

You would notice fuel on the road , in the dark, just round a corner, really?
Its those sorts of situations its more of an issue, and where it normally happens, round a corner as lorries have overfilled tanks