Brexit, for once some facts.

[oldgroaner]

I, ve wondered where we now stand with this new technology for such as braking for dogs in road. I was always told don't risk those behind (or yourself) braking for an animal. ie)Check mirror before braking for a stupid mutt... Emergency stopping with a 30 tonne truck behind you is not exactly safest thing to do... If it's person, we'll no choice.. But don't think radar or any current technology can differentiate...
Can you imagine outcry. Car braked to save some mongrel and caused mayhem.. Meanwhile dog runs off????

You are so right on this one! there are many unappreciated hazards the designers didn't accoount for, including the level of distraction the driver can face especially in close quarter encounters in narrow lanes.
And I forgot to add if your phone is linked in that overrides both the Navigation spoken commands and masks the sounds of the proximity detectors.

Plus possibly the driver could become too reliant of the automatic safety the system offers (conditional to conditions being favourable)
Apparently this has been brought up in Parliament too already.

 

[flecc]

cameras and all types of sensors get better all the time.

Making driverless even more impossible!

Even now they are suffering from false detections. Once full reliable detection for all circumstances is achieved, the cars will be paralysed by false detections.
.

 

[oldgroaner]

https://dailym.ai/2JGiaGJ

Corbyn getting support in DM? What is the world coming to.

I read that and wondered the same thing....
Decided I was seeing things and carried on.....

 

[flecc]

I read that and wondered the same thing....
Decided I was seeing things and carried on.....

They're just waking up to the fact that despite all their mud slinging, Corbyn is still a real threat.

So showing that is now a better tactic to alert their true blue readers to the danger if they become complacent.
.

 

[oyster]

that's the same situation with all kinds of computer based systems in daily use. If your microwave's clock goes wrong, you may set fire to your house etc.

the datastream from an incoming car includes the transponder ID, the reference system eg WGS1984, satellites identification and readings. They are normally compared to your car's own datastream for fixing. Your car can then look up its databases for the model of the incoming car, uses its imaging system to identify the incoming car among all the objects that your car's camera(s) can see. Galileo will be certainly useful for driverless cars.

Your microwave doesn't have to rely on getting its time from the microwave in the house over the street. It is internal, it is not likely to be affected by bad data. Yes, of course it can go wrong. That is why it, and much else, advises not to be used unsupervised. (Probably - many do but some might not.)

Could you give me an estimate as to how long it would take for your vehicle to gather all the information you are suggesting, for all vehicles around it, decide how to estimate for vehicles which have not yet had the devices installed (or are broken), and then decide what to do?

Indeed, just how many vehicles are relevant? - take heavy but moving traffic on the M25 for an example. At closing speeds of, say, 140 mph, the number of vehicles coming into, and leaving, the zone of relevance could be pretty substantial. You can't just ignore the other carriageway, things happening there could end up being of importance.

 

[oyster]

My conclusion is, why did they bother?

I agree. I do like adaptive cruise control - very largely because it can make my right foot more comfortable AND avoids me having to keep looking at the speedo. But I would reject a partial system where I have to go from being totally side-lined to being in control in milliseconds. It is precisely because I am still the one in charge that I am comfortable with it. Even then, I wonder how someone new to driving would cope? I have got an immense number of totally self-driven miles under my belt and that experience counts.

 

[Woosh]

Your microwave doesn't have to rely on getting its time from the microwave in the house over the street.

eg you set the microwave on for 2 minutes, if it does not stop, it'll burn your food.

 

[Woosh]

Could you give me an estimate as to how long it would take for your vehicle to gather all the information you are suggesting, for all vehicles around it, decide how to estimate for vehicles which have not yet had the devices installed (or are broken), and then decide what to do?

it's about 10 milliseconds at present.

 

[Woosh]

Indeed, just how many vehicles are relevant? - take heavy but moving traffic on the M25 for an example. At closing speeds of, say, 140 mph, the number of vehicles coming into, and leaving, the zone of relevance could be pretty substantial. You can't just ignore the other carriageway, things happening there could end up being of importance.

this image gives an idea the field of vision for a driverless car:

 

[oyster]

it's about 10 milliseconds at present.

Surely, it is simply not possible at present?

I sincerely doubt 10 milliseconds will be achievable in the real world, especially if any data needs to be retrieved from external sources. (What? You've never heard of a 2019 Audi A1? Ah - it's just out for a bit of road testing.)

If you are going to rely on satellite information, let us consider how to handle tunnels, multi-storey car parks, even roads lined with tall buildings. We need to implement that before we can really get started.

We need to know how thunderstorms might affect communications.

And on and on...

 

[oyster]

eg you set the microwave on for 2 minutes, if it does not stop, it'll burn your food.

Which is why it should not be used unattended.

 

[Woosh]

Surely, it is simply not possible at present?

I sincerely doubt 10 milliseconds will be achievable in the real world, especially if any data needs to be retrieved from external sources. (What? You've never heard of a 2019 Audi A1? Ah - it's just out for a bit of road testing.)

If you are going to rely on satellite information, let us consider how to handle tunnels, multi-storey car parks, even roads lined with tall buildings. We need to implement that before we can really get started.

We need to know how thunderstorms might affect communications.

And on and on...

all the objects are detected as your vehicule sees the landscape and already registered in the driving database. The 10ms is enough to update its database with the changes and comes to its decision.
Speed is not the problem, we have plenty of computing power. What we don't have is the artificial intelligence software.
We can't detect 3D objects with sufficient accuracy at the moment.

 

[oldtom]

This piece about American society can be equally applied to the UK:



Tom

 

[flecc]

What we don't have is the artificial intelligence software.

And there's another crunch, as better brains than mine have observed, we haven't even begun to create intelligence yet. So called artificial intelligence is nothing of the sort, it's simply better and better calculators.

Determining whether to brake or not for the object passing in front in any given cicumstance is only in the province of human intelligence and knowledge.

Never for the paper bag or similar.

Never for the balloon someone has tied to a lamp post and which is fluttering in the breeze.

Sometimes for the youngster's ball that's run into the road, if safe to do so.

Always for the pedestrian on the pavement who isn't in detection line but having tripped, appears just about to fall off the edge of the kerb.

Never for the squirrel that's run into the road, but sometimes for the cat or dog if safe to do so.

Never for the low flying blackbird in Spring.

And so on. Detection problems will eventually be solved, but at the cost of too many false detections to be usable and still with the inability to choose what to do in each event according to circumstances.

The people developing this have blinkers on giving tunnel vision. Being so obsessed with the engineering possibilities, they don't seem able to see the impossibilities in the real world when without human intelligence.
.

 

[oldtom]

Some say it'll be great to regain British sovereignty and self-determination. Then we can restore all power to the monarch and disband parliament. Surely that would work?



.......maybe not!

Tom

 

[oldtom]

I liked James O'Brien's tweet:


Tom

 

[Woosh]

Detection problems will eventually be solved, but at the cost of too many false detections to be usable and still with the inability to choose what to do in each event according to circumstances.

artificial intelligence is not yet intelligence at present but it is progressing fast. We have made good progress in agriculture. Starting with driverless plough tractors a few years ago, now we have robots to pick fruits and plant seedlings. You see the progress to level 4 driverless with very few casualties. If you look into previous steps: road markings, traffic lights management, the airbags, then automatic braking and car parking etc - progress is rapid.
I am aware that the industry is full of optimistic people wanting to promote their business, I know it will take longer than predictions, just look at nuclear fusion. It will take as long as needed, the demand is there, waiting. Level 5 will be a very important step in helping the old and disabled to have a greater autonomy.

current prediction: level 4/5 on the roads: 2020/1, level 5 urban driving: 2030.

 

[oyster]

Never for the low flying blackbird in Spring.

About thirty years ago, I was driving on a lovely sunny afternoon, along a dual carriageway with excellent visibility, and hardly any traffic. Saw some ducks over a field, they flew across the road, and I really thought nothing of it. Until the last one looked much lower - and suddenly it hit my windscreen just behind the rear-view mirror. The surprise, and the very loud bang, could well have caused an accident in other circumstances. Got a crack in the screen but could very easily have shattered it.

If I had realised earlier, almost certainly I could have slowed down sufficiently to let it get by. Either my judgement was bad - or it did fly lower than its original trajectory had indicated. Either way, it should be taken into account!

 

[flecc]

current prediction: level 4/5 on the roads: 2020/1,

No, and I should be around to say I told you so. Level 4 restricted to very few locations rather later than that perhaps. Even then they'll often be a considerable nuisance to the great majority still driving their cars due to the unexpected things they'll do from time to time.

I think level 4 is potentially very dangerous, relaxing drivers sufficiently for them to be unable to assume control quickly enough when called for. Goodness knows enough drivers fall asleep or doze even when doing all the driving.

level 5 urban driving: 2030.

One day in certain dedicated locations, but lots of town and city driving will always be beyond it unless we radically change our streets, overhaul what we all do and have far greater urban discipline. That's all unlikely in any forseeable future when we can't even repair potholes. Also the predictions for no steering wheel or even no manual controls are bunkum. I've previously posted just two of the situations why that cannot ever happen, and there are very many more.
.

 

[Zlatan]

No, and I should be around to say I told you so. Level 4 restricted to very few locations rather later than that perhaps. Even then they'll often be a considerable nuisance to the great majority still driving their cars due to the unexpected things they'll do from time to time.

I think level 4 is potentially very dangerous, relaxing drivers sufficiently for them to be unable to assume control quickly enough when called for. Goodness knows enough drivers fall asleep or doze even when doing all the driving.



One day in certain dedicated locations, but lots of town and city driving will always be beyond it unless we radically change our streets, overhaul what we all do and have far greater urban discipline. That's all unlikely in any forseeable future when we can't even repair potholes. Also the predictions for no steering wheel or even no manual controls are bunkum. I've previously posted just two of the situations why that cannot ever happen, and there are very many more.
.

Agreed totally and even if we could solve all the issues isn't the point of the car to take somebody somewhere. Can't quite see how taking nobody somewhere is beneficial.