It amuses me that this isn't even a sentence and appears to be utterly irrelevant to my post. Classic.

Yes it is a sentence response to the term captain obvious (ie tell that). Now how about some actual contribution to the thread. I bet writing code to do all these things in racing mode is a decent research team project in some engineering departments today.

I like how you're bad at grammar. And that you're hyper-defensive all the time. That also amuses me.

My contribution was resolving OP's stated question in less than 100 words. But it's true that I probably could have done it more efficiently. Let me do that now.

False.

Consider spinning in a circle with a very long rigid rod. Do you think the Gs near the pivot point will be the same as the Gs at the very end of the rod?

That is all great but to someone asking the question you still need to show the connection is the same angular velocity for both ends. That is trivial to one that knows about rigid bodies etc but then these guys wouldn't be asking the question with the same times of flight condition to begin with. My posts are long so that people that care to read can gather maximum information. You can use the sling - ancient weapon equivalent too and now that you have seen the discussion on angular velocity can understand what you know about it already.

https://en.wikipedia.org/wiki/Sling_(weapon)

(eg imagine slings of different length etc from 5 cm to 0.5 meter or use your cable mouse as example lol)

I bet >90% of the people reading the thread havent seen in their lives the general derivation of the centripetal acceleration either including some high school teachers of math and science. Some good people studying calculus recently can see their vector analysis in action too.

Thanks!

What's your over/under for when the first fully self driving car wins the Daytona 500?

PairTheBoard

I dont know but i guess within 5 years they can be topping humans in lap qualification terms.





they may top us in test obstacles etc.

http://www.telegraph.co.uk/news/scie...irst-time.html



Teaser for the future?

Never. I don't think they will ever allow the cars to compete with humans like that. But there will likely be self-driving races eventually.

Sometimes when I get behind on work or some other form of psychological stress, I'll have this recurring dream where I'm driving a car and simultaneously controlling a car remotely in front of me. Each time I end up lagging behind the remote car due to traffic to the point I can't see it and it's quite scary because I know it's gonna crash into someone.

AI doesn't rely on a "proper" theoretical model of the car motion. It reaches the conclusion by a half-way decent model followed up with real-world experimentation to reach the best possible solution.

You don't even need the model to consider massively important things, such as torque-steer, since it is corrected for during experimentation.

They have been doing this for years, fwiw.

Agreed. I could easily create a better running-back than is currently available using machines. A very grabby autonomous tank would work pretty well. Probably not going to be in the NFL anytime soon.

Autonomous racecars would be boring as ****, also.

You should try to have better dreams. When I am stressed, I have dreams about 4-somes and 5-somes. The good kind of x-some, of course, but it is generally a logistical nightmare.

And tell me again why you quote this and not the paragraph above that which talks about what your criticism implied?


"The theoretical is more complex in general than i presented because the car itself may have different placement/balancing of tires and different tire properties from tire to tire plus there are others small corrections here and there but in general it wont be that bad if an AI could push the limits of the theory with constant experimentation to learn the car properties because ideally you want to push the car to the real limits of tolerance of sliding and pick the curve that allows to have as much as possible higher speed everywhere while feeling maximal static sideways friction everywhere it cannot be avoided (or say a little below that to be confident to not fail)."

In this site i am trolled 24/7 and one day i will troll back and never post again.

There were no implications. I said what I said pretty darn directly.

The implication and light sarcasm is always there present as if i was born yesterday and dont know that real world is very complicated and cannot be all equations or that some people like to only superficially understand how things work because it is enough for their purposes.

I am not gonna tell a mechanic or a race car driver their job here. But you better believe it if the mechanic and car manufacturer gives me exact details about how their equipment functions i can offer ideas to them what to look for too.

Proper research uses both theory and experiment and in this case some likely solution only requires light acceleration or deceleration and mostly steering to define the path. The actual process that accelerates the car is the friction from the road that is manipulated by proper torque from the engines so detailed knowledge of the torques and the response of tires and ground is going to go into the equations and be further fine tuned by experiment.

You may go full brutal theory on it or just fit it to experiment. A combination of both helps do a better work always. Someone has to guide the experiments where to go (if a clean albeit complex modeling exists) and what to test and how with new ideas and those will be mathematically inspired often and then made to fit to what actually is available technologically. If you find an ideal curve you then can look how to manipulate the car brakes/gas/steering to arrive at that curve as close as possible by timing your choices. That timing is beyond human control. It has to be very precise hence why AI may have an edge eventually. Experimentation connects the physics with the response of the equipment you have. You cannot know the connection without detailed experimentation of these functions.

I wont be surprised if ideal curve finding is still an open topic for research with room for improvements.

That is, unless you're actually driving. I really don't know how people stand to watch cars driving in circles for 500 laps.

I can see a 5-20 lap race holding my attention, I think.

What's with all the hate? Feels like Politics up in here.

This squirrel has no testicles.

But if it did...

I like that you've got some fight in you.

https://www.youtube.com/watch?v=ltwJ9H7rpoM#t=121

Actually what you need is to talk to me live and then i want you to tell me there that a decent human being would say to someone that they knew for years "let today be the day".

Because i would never do that to you.

The squirrel however cut the ropes i use outside to hang hard to dry big items like blankets that do not fit well in washers and dryers so he would get the punch for sure to get into orbit. The bastard cut the ropes and then left and then did it again with new ropes too. Its not as if he cut them and ate them or took them to play. He is a real trolling bastard!

Have no doubt that my true version of trolling is indeed to log off for real. Its the only trolling that would come out of me.

People deserve to be in communities that they interact experiencing respect and friendship. One of the most important things in a discussion is how you choose to disagree with others. Life is not waiting for us indefinitely to be spent with people that do not care enough.

I can only hope people that do not post here lately or for a long time now are somewhere out there happily moving on and not in any trouble.

Again, what I said was very direct.

Maybe not at Daytona but I suspect there will be some race somewhere in the world that will allow it, at least experimentally, for the publicity. Some race that's looking for the extra attention.

PairTheBoard

The humans would have to apply game theory vs the top AI car and block it letting another human driven car run a distance of safety in the meantime ahead, even cause unlucky accidents as needed. Then the AI would have to try to outplay the collusion provided it understands it.

Oh the fun it will be in the coming decades. Not only it will happen, it will be advertised as the battle of the ages. And in the it will be exactly that.

Imagine the machine coming out and saying hey guys you try to collude and i will try to outplay you without destroying you or harming you because its a better challenge just to see that i am ready for what is coming and you are not.


I always come back to the 2 best AI vs human ads in history and will shamelessly link them forever for anyone that havent seen them any time the topic gets close enough lol;

first it was



and then




The car equivalent ad will be glorious.

I'd say it to you to your face, but you'd also hear the vocal inflection of it and laugh.

You open yourself up to being trolled because you play that "woe is me - I'm the victim" line waaaay too often. You also come off as stupidly arrogant and kind of intellectually inept in the way that you take a position and appear to have absolutely no desire to listen and change your opinion about anything. That's also easy fodder to mess with.

As soon as you get your head out of your rear end and start interacting with people in a way that actually resembles what a real conversation looks like (not the endless ramblings of someone who believes he knows more than everyone else in every single subject, and of someone who writes 10,000 words but can't even answer the question being asked), you'll find the level of trolling will diminish. But until then...