F1 To Be Decided By Wins

[Quiet One 15]

You don't need wires. It is perfectly possible for a car to know exactly where it is on track using positional systems. Mauron son of Sauron would have told you this if he had a clue.

Well...it was just an idea.

Navigation? Motion planning? FFS stop googling to pretend you have a clue. It is a one way round track. You need n navigation, or motion planning. A rudimentary 'don't crash into cars around you' algorithm would be a great start.

Both approachs are possible.

Indeed. An AI driver would have caught the signs of a car going out of control far earlier and responded instantaneously. No human can compete with that.

Yup. And nobody cares if the car crashes, anyways. Well, except the owner, of course.

[mikathegreat2 2]

What's surprising the crap out of me is that maure has'nt mentioned "Lewisterics" yet !!!

[Jean Todt 4]

Here.

http://www.youtube.com/watch?v=AO5zsE0jLU0

[Jean Todt 4]

Here.

http://www.youtube.com/watch?v=AO5zsE0jLU0

[maure 1]

Budgets of millions have been spent on designing AI racecars?? Where?

They dont need to make sense of their surroundings. They are driving in a small well defined environment. Who cares about the surroundings, as long as the car knows where it is on track.

bulls##t. The DGC is not relevant here, it has cars negotiating traffic rules, urban environments etc. The problem of driving an F1 car around a track is much simpler.

You are not half as fun as Murray.

1. Millions have been spent on driverless vehicle research.

2. Your _claim_ that "they dont need to make sense of their surrounding" is your reduction of the problem to absurdity. After all, part of the surroundings are other cars.

3. DGC is relevant as is all research in the area. Besides, your limitation of DGC to "urban environments" means you don't know much about DGC.

4. The problem of driving a vehicle at 300 k/h is much much much more complicated than DGC and, at this point, impossible... except, of course, to the two resident geniuses that claim is all easy but are _unable_ to give any details whatsoever on how to accomplish it.

[maure 1]

What's surprising the crap out of me is that maure has'nt mentioned "Lewisterics" yet !!!

Go on, say something that merits it me calling you a lewisteric... try me.

[maure 1]

Ok, I see the point. I was talking about merely making some cars race with no driver. That can't be so much harder than a slot car race (and almost as entertaining). You are talking about something more ambitious. Well, if they ever decided to go the AI way, I bet they will choose my "dumb cars running in circles" path over your "cars with no driver acting almost human".

But yes, for cars to act with such a degree of precision, you would need an almost human brain. If they can't solve the "true vision" AI problem, I don't think we will see anything too complex anytime soon.

Yes, if the car malfunctions, it will become a missile. Much like today, only it won't carry a likely victim inside. Adequate safety measures would not be more complex than today. They should be easier, if any. Bigger runoff areas, more space between the public and the cars and that's it. No ambulances, no medics.

Kubica's crash last year was a great example of a car at maximum speed totally out of control. The only persons in danger were the drivers and the track personnel (if there were some of them around). It wasn't too bad. I can't see why it would be worse without humans involved.

As a parallel competition? Mmnmmh...I agree. They are trying to make some robots play football, after all. Still, my own preferences go towards driver controlled cars. But yes, it would be interesting to watch them try!

In order to be interesting to me, a driverless F1 competition requires fully autonomous cars. Otherwise, it is just one of those slot-car games (scalextric?) but scaled up.

But even if you were to go for the toy-like competition cavallino speaks of, you still have many other problems to resolve. F1 is extreme in its consideration of influences (or variables) and an AI would have to be able to respond, for example, to changing track conditions, the so-called "feel" that drivers get off the tarmac... Ultimately, to steal away from the role of the driver to justify its replacement with a bot is an absurd proposition. Might as well look into remote control cars...

[Quiet One 15]

In order to be interesting to me, a driverless F1 competition requires fully autonomous cars. Otherwise, it is just one of those slot-car games (scalextric?) but scaled up.

But even if you were to go for the toy-like competition cavallino speaks of, you still have many other problems to resolve. F1 is extreme in its consideration of influences (or variables) and an AI would have to be able to respond, for example, to changing track conditions, the so-called "feel" that drivers get off the tarmac... Ultimately, to steal away from the role of the driver to justify its replacement with a bot is an absurd proposition. Might as well look into remote control cars...

My sentiments, exactly.

My only doubt resides on how different it would be, from the spectator's point of view, between the scalextric approach and the "true AI" approach. My guess is that it would not make for much of a spectacle, but then again, I am a fanboy

I don't confuse drivers with starlets, but I love the fact that they are human with their quirky personalities and all that. I wouldn't want it any other way.

[Max Mosley 2]

Yes it would be duck as full without the drivers. For me they're the whole point. But this would at least be some fancy research that I wouldn't ridicule, rather than the usual guff we get about innovations that your dad could probably have put on your tricycle himself. Pumping tyres up with something other than regular air causes a spy scandal, I mean come on!

I can't believe it's really as hard as Maure makes out. He's never been right about anything so far, and that's good enough logic for me.

Btw I was thinking today about cats. No doubt the computational power of a cat brain is orders of magnitude greater than any artificial computer yet made. So why is that brainpower necessary and what does a cat do with it? You wouldn't think the life of a cat was so complicated...

[monza gorilla 1]

I can't believe it's really as hard as Maure makes out. He's never been right about anything so far, and that's good enough logic for me.

Btw I was thinking today about cats. No doubt the computational power of a cat brain is orders of magnitude greater than any artificial computer yet made. So why is that brainpower necessary and what does a cat do with it? You wouldn't think the life of a cat was so complicated...

I would guess that a great portion of the small cat brain is used in spatial computation. You gotta catch them meeces somehow. Also the highly sensitive olfactory and auditory capabilities.

I think Maure has hit the nail on the head in one respect, though. Feel. But, then again, why would an AI need that? Wouldn't yaw sensors and the like be able to detect the limits of adhesion? Or would it be like my Beemer, and mess with your head by cutting the power just as you're beginning to have fun. I suppose that, as you and Cav say, it wouldn't be difficult to get a single AI car lapping at some fairly high speeds. But in a race situation? I think that's probably a tall order at present. I speak, of course, from a position of unparalleled ignorance.

[Quiet One 15]

Yes it would be duck as full without the drivers. For me they're the whole point. But this would at least be some fancy research that I wouldn't ridicule, rather than the usual guff we get about innovations that your dad could probably have put on your tricycle himself. Pumping tyres up with something other than regular air causes a spy scandal, I mean come on!

You have serious psychological problems with innovation anywhere outside your University

I can't believe it's really as hard as Maure makes out. He's never been right about anything so far, and that's good enough logic for me.

Well, a "true AI" would be really hard. As I mentioned in a previous post, you must deal with things like "the true vision problem" that is, developing a computer system that can "see" the way the humans do. So far it remains unsolved and there are serious debates over whether it can be solved or not in practice.

Btw I was thinking today about cats. No doubt the computational power of a cat brain is orders of magnitude greater than any artificial computer yet made. So why is that brainpower necessary and what does a cat do with it? You wouldn't think the life of a cat was so complicated...

Too much time in your hands...

[pumpdoc 2]

2 plus pages of wasted space.............................................

[cavallino 2]

I think Maure has hit the nail on the head in one respect, though. Feel. But, then again, why would an AI need that?

Exactly. You can detect exact position on the track corresponding to an embedded track map, velocity, orientation, grip/ wheelspin/ traction on each wheel, lateral and un-lateral acceleration. you dont need feel.

[cavallino 2]

You have serious psychological problems with innovation anywhere outside your University

Well, a "true AI" would be really hard. As I mentioned in a previous post, you must deal with things like "the true vision problem" that is, developing a computer system that can "see" the way the humans do. So far it remains unsolved and there are serious debates over whether it can be solved or not in practice.

AAAARRRRRRGGHHHHHH

You don't need true AI you don't need computer vision how is that at all relevant. it's like saying a chess playing program needs to be able to look at a chessboard and figure out the position by pointlessly using image processing algorithms. All you need is an adequate AI, that's how you solve AI problems. As maure would know if he had a clue. Maure wolud look at creating a chess playing AI as 'creating a humanoid who can move pieces on a real chess board and read the board postion off it - an utter waste of effort'.

[cavallino 2]

1. Millions have been spent on driverless vehicle research.

And?

2. Your _claim_ that "they dont need to make sense of their surrounding" is your reduction of the problem to absurdity. After all, part of the surroundings are other cars.

It is all about reduction. That is how you solve AI problems, by removing the superfluous. Reduction is the essence of it. And that is why you have no clue what you are talking about.

Take the surroundings. Do the cars need to 'see' oher cars? No. They just need to be aware if another car is close. That is all. No since the only thing that can be close is a car, we reduce the problem. The car needs to be aware if another object is close and the speed it is moving at. That's radar. They just need radar. Not bloody make sense of their surroundings.

3. DGC is relevant as is all research in the area.

Not as much as you pretend it is.

Besides, your limitation of DGC to "urban environments" means you don't know much about DGC.

They're wprking with more complex systems, which have more unknows / unpredictable scenarios.

4. The problem of driving a vehicle at 300 k/h is much much much more complicated than DGC and, at this point, impossible... except, of course, to the two resident geniuses that claim is all easy but are _unable_ to give any details whatsoever on how to accomplish it.

Rubbish. The speed is not the most important thing. An F1 car at 200 mph has far less safety issues than a car doing 40mph on a public road. A car on a public road needs to always avoid pedestrains. It needs to read traffic lights with zero error - missing one light could kill someone. Not seeing a pedestrain could kill someone, straying out of your lane could kill someoene. F1 track - none of these issues matter. So it's not like '300 kph omg that sounds so haaarrdd lololololol'

But even if you were to go for the toy-like competition cavallino speaks of, you still have many other problems to resolve. F1 is extreme in its consideration of influences (or variables) and an AI would have to be able to respond, for example, to changing track conditions,

Yes you bloody idiot a car can read changing track conditions far better than a driver.

the so-called "feel" that drivers get off the tarmac...

A driver with traction control will always beat one without out of the corner if they have the same entry. All the bloody feel in the world can't beat the electronics.

Ultimately, to steal away from the role of the driver to justify its replacement with a bot is an absurd proposition. Might as well look into remote control cars...

Tjhat's not the point here. The point is you are a pretentious lying ignorant clueless idiot, and I can't believe anyone actually believes anything you write. I mean, you are so incredibly staggeringly wrong all the time. You'r like the most scary possible result of the monkeys with typewriters experiment - a monkey who can type crap but which is hard to tell apart from real crap. What you post is no better than a monkey on a type writer, but it's far more of a paint to prove that it is no better.

[Max Mosley 2]

I would guess that a great portion of the small cat brain is used in spatial computation. You gotta catch them meeces somehow. Also the highly sensitive olfactory and auditory capabilities.

Yes, I guess it must be something like that. It seems amazing to me that a cat might have more computing power at its disposal than every human-built computer in the world put together.

You have serious psychological problems...

Too much time in your hands...

Guilty on both counts. But then that's to be expected in a university.

2 plus pages of wasted space.............................................

And that's just Maure's verbiage.

Maure wolud look at creating a chess playing AI as 'creating a humanoid who can move pieces on a real chess board and read the board postion off it - an utter waste of effort'.

And Maure's AI cars would have to whine, cheat and steal like Alonso whilst being as humble and magnanimous in victory as Lewis. All on their own.

[Quiet One 15]

AAAARRRRRRGGHHHHHH

You don't need true AI you don't need computer vision how is that at all relevant. it's like saying a chess playing program needs to be able to look at a chessboard and figure out the position by pointlessly using image processing algorithms. All you need is an adequate AI, that's how you solve AI problems. As maure would know if he had a clue. Maure wolud look at creating a chess playing AI as 'creating a humanoid who can move pieces on a real chess board and read the board postion off it - an utter waste of effort'.

Ok, you are confusing "must" with "would be nice if"

Nobody says you "must" have complex AI to make some cars go around a track. But it would be nice ("nice" here meaning "being fun for a couple of Discovery Channel's half hour programs) to see how humanlike can you make those cars race.

I've read somwhere that, in fact, they were trying to make robots "watch" a chessboard and play chess like that. Why do that? Just because of the challenge, and because it lets you understand more about AI if you make a robot that can watch the chessboard and pick the right move than a computer that merely runs through every conceivable combination. Again, is something "nice to do", not something you "must" do in order to win a game. The game (or car race) are just excuses.

Other than that, I still think it will be too boring to last more than a couple of races.

[mikathegreat2 2]

Go on, say something that merits it me calling you a lewisteric... try me.

Unlike Fed Up (no offence), I don't suck hi willy every day even if he's the biggest pr1ck ever! I like Lewis for his driving ability & his aggresion (even if it does get a lil bit out of hand). What I think of your behaviour about "Lewisterics" is childish & I don't unserstand why you are obssesed with them??

[maure 1]

My sentiments, exactly.

My only doubt resides on how different it would be, from the spectator's point of view, between the scalextric approach and the "true AI" approach. My guess is that it would not make for much of a spectacle, but then again, I am a fanboy

I don't confuse drivers with starlets, but I love the fact that they are human with their quirky personalities and all that. I wouldn't want it any other way.

I can tell you from my experience with AI's that, in most cases (not always though), the most fun is the designer's. If you know enough about programming to write up a simple game, you should be able to develop a simple AI to play the game on its own. The basics of AI are not complicated for very simple problems and the experience is quite remarkable. As you learn about the game, you program the AI with those strategies and see it resolve (on its own) novel scenarios with those strategies. If you are then able to make the AI report its "thinking process" as it is progressing,... you'll get a kick out of it. I'm sure.

You bring up the issue of quirks and that, actually, is part of AI design. As strategies are built into a knowledge database, these reflect certain choices on the part of the developer. Sure, it is possible to go the artificial neural network route and this requires no initial state (any random state works) but training of the ANN will be influenced by the developer (learning parameters). Ultimately, the behavior of any AI is influenced by the designer. So, instead of the quirks of drivers, we would have the quirks of of developers... another dimension of quirk, I guess.

[maure 1]

And?

It is all about reduction. That is how you solve AI problems, by removing the superfluous. Reduction is the essence of it. And that is why you have no clue what you are talking about.

Take the surroundings. Do the cars need to 'see' oher cars? No. They just need to be aware if another car is close. That is all. No since the only thing that can be close is a car, we reduce the problem. The car needs to be aware if another object is close and the speed it is moving at. That's radar. They just need radar. Not bloody make sense of their surroundings.

Not as much as you pretend it is.

They're wprking with more complex systems, which have more unknows / unpredictable scenarios.

Rubbish. The speed is not the most important thing. An F1 car at 200 mph has far less safety issues than a car doing 40mph on a public road. A car on a public road needs to always avoid pedestrains. It needs to read traffic lights with zero error - missing one light could kill someone. Not seeing a pedestrain could kill someone, straying out of your lane could kill someoene. F1 track - none of these issues matter. So it's not like '300 kph omg that sounds so haaarrdd lololololol'

Yes you bloody idiot a car can read changing track conditions far better than a driver.

A driver with traction control will always beat one without out of the corner if they have the same entry. All the bloody feel in the world can't beat the electronics.

Tjhat's not the point here. The point is you are a pretentious lying ignorant clueless idiot, and I can't believe anyone actually believes anything you write. I mean, you are so incredibly staggeringly wrong all the time. You'r like the most scary possible result of the monkeys with typewriters experiment - a monkey who can type crap but which is hard to tell apart from real crap. What you post is no better than a monkey on a type writer, but it's far more of a paint to prove that it is no better.

The usual contradictory rambilngs from cavallino.

First he says that no money is being spent on research on the subject, then that the subject is other, then that the problem must be reduced, then that the problem is other, then...

Should I even bother? Your ignorance is so profound that you confuse the reduction of a problem into manageable components with the reduction of a problem in a manner that changes, even eliminates, the problem.

The only amusing thing is that you consider yourself a "pretentious lying ignorant clueless idiot". It could be no other way. After all, only one such person can consider problems as earth's climate or AI a "simple" matter where all is known...

[maure 1]

Unlike Fed Up (no offence), I don't suck hi willy every day even if he's the biggest pr1ck ever! I like Lewis for his driving ability & his aggresion (even if it does get a lil bit out of hand). What I think of your behaviour about "Lewisterics" is childish & I don't unserstand why you are obssesed with them??

So you like Hamilton, so what? So did I once and, in fact, still enjoy some of his driving. Again, so what?

Now insult drivers other than Hamilton to cover up, for example, for Hamilton insulting Kimi, and I will call you a lewisteric because that is what you would be. Is that obsession? Nope. In that case, it would be a fact that you are acting like a lewisteric.

If the behavior persists and you continue to insult, particularly if you use bigotry and hatemongering as Murray, Fedup, Abbas, etc have in the past, you would move from "acting" like a lewisteric to _being_ one.

Perhaps it would help you understand if I tell you that I would much prefer for Murray to be who he claims to be, that is, a world-reknown expert on several academic fields (which is the implication of his claims... even if he does not even comprehend that) and that same applies to cavallino. Why? Because that would make them interesting. Instead they run their mouths, making fools of themselves with grand showings of ignorance, which is not interesting but is amusing. Interesting is much better than amusing, but amusing is what this forum yields... in general.

[maure 1]

I would guess that a great portion of the small cat brain is used in spatial computation. You gotta catch them meeces somehow. Also the highly sensitive olfactory and auditory capabilities.

I think Maure has hit the nail on the head in one respect, though. Feel. But, then again, why would an AI need that? Wouldn't yaw sensors and the like be able to detect the limits of adhesion? Or would it be like my Beemer, and mess with your head by cutting the power just as you're beginning to have fun. I suppose that, as you and Cav say, it wouldn't be difficult to get a single AI car lapping at some fairly high speeds. But in a race situation? I think that's probably a tall order at present. I speak, of course, from a position of unparalleled ignorance.

Be careful now, don't say I might be right.... Although, I grant you, Murray's expert knowledge of logic does not include understanding that a single contradicting case is all that is necessary to bring down an "always" (12yo math where I come from)... that's what makes him so "right" when he is so desperately out of ideas.

But, hey, cavallino has just flipped out rainbow-like and Murray is now nothing more than excuses. It can only mean that the two resident geniuses are about to publish "seminal" papers on "reputable" journals that will reveal to the world why it is _all_ that easy...

[BradSpeedMan 6]

Rubbish. The speed is not the most important thing. An F1 car at 200 mph has far less safety issues than a car doing 40mph on a public road. A car on a public road needs to always avoid pedestrains. It needs to read traffic lights with zero error - missing one light could kill someone. Not seeing a pedestrain could kill someone, straying out of your lane could kill someoene. F1 track - none of these issues matter. So it's not like '300 kph omg that sounds so haaarrdd lololololol'

interesting theory

[Grabthaw the Hammerslayer 4]

The thing about AI is that it requires humans to create it (currently) and unless you have a superfast AI that can evolve (which we don't yet) someone has to input all the parameters and scenarios that *could* happen. Which leaves a problem. Let me illustrate:

Paris air show a few years ago where the Airbus A320 was being shown off. This was one of the early fly by wire planes. The pilot flies past at almost stall speeds to demonstrate the clever computer control. Goes to pull out, throttles on max. Plane carries on and crashes in a wood. Subsequent investigation reveals that the computer over-rode the pilot as it thought that the pilot's actions went beyond the expected limits of what the plane could/should do.

This is where the problem occurs. I remember the race when Hakkinen hit a groundhog (Canada) or when an Irish guy ran across the track with a flag. Humans have the capability to recognise situations and react accordingly. OK we don't react as quickly as computers but we are more flexible in terms of how we identify situations, analyse and react. I don't believe AI is there yet. I've seen some of the stuff going on with driverless cars and its very interesting - clearly progress being made, but I wonder how an AI car would deal with the situation like when Alonso crashed on the home straight in Brazil a few years back, spinning across the track, debris everywhere. Or like when Schui drove the best part of a race in Silverstone stuck in 5th gear or in those situations where team orders come into play and one car has to give way or making a decision to switch to dry tyres earlier in a drying track scenario to gain competitive advantage, or when someone makes a mistake and lets someone else through, or.... (I could go on).

Yes of course, some boffin could create something to do this, but it is not there yet. I predict a crash-fest extraordinaire.

Moreover whats the point? Is that going to be exciting? No! F1 is about man and machine working together and what makes it interesting are all the flaws and f***-ups. I'd want to see robotic F1 about as much as CGI generated porn......

[bmwf1guy 0]

Bernie doesn't seem to have the support he says he does. Remember that he said that the teams are all in favor of the medal system. Turns out that no team has sat down to discuss their position on this yet. Bernie caught in another lie.

I don't like this system. In a series, or league such as football, hockey, rugby or whatever, these all run on points because they are series just like F1. Amateur sports where it is one event such as the Olympics, medals work then. I don't see the current system being broke so I don't want it fixed.