PDA

View Full Version : Rigid Body Dynamics is an amazing topic!



WhatEver
04-06-2001, 02:30 PM
I'm in awe at the things some people can do with programming, and Rigid Body Dynamics is one of them. I think it's so cool that computers have aloud us to take physics equasions and put them into a real time visual demonstration.

How soon do you guys think it'll be when these physics will be implimented in games involving human bodies? Or have they already?

Rigid Body Dynamics is the most in deapth topic I've ever seen. I guess this topic will have to be put on my to-learn list.

Nutty
04-07-2001, 12:54 PM
Got any beginners links for Rigid Body stuff?

cheers,

Nutty

iss
04-07-2001, 01:13 PM
Rigid body dynamics is a large topic. Newcomers to computer graphics think that this is a new topic. Actually, the mathematics for much of this was worked out over 100 years ago as Classical Mechanics as a branch of physics. Be forewarned, the math is heavy.

The best introduction I know of is at: www.cs.cmu.edu/~baraff/sigcourse/index.html (http://www.cs.cmu.edu/~baraff/sigcourse/index.html)
Also look at: www.cs.cmu.edu/afs/cs.cmu.edu/user/baraff/www/pbm/pbm.html/ (http://www.cs.cmu.edu/afs/cs.cmu.edu/user/baraff/www/pbm/pbm.html/) . The math level is at senior undergraduate or first year graduate physics/engineering/mathematics level.

It is quite an undertaking (usually as a PhD thesis, literally ) to become an expert in this topic (i.e. develop new algorithms). If you like math, Goldstein’s Classical Mechanics is an excellent text. Personally, I prefer quaternion over matrix rotations but this is akin to the EMACS vs VI war. (EMACS is better, right?)

One of the most interesting applications is combining Classical Rigid body Mechanics with Continuum Mechanics to do Deformable modeling. Demetri Terzopoulos www.cs.toronto.edu/~dt/ (http://www.cs.toronto.edu/~dt/) is one of the pioneers this topic.


[This message has been edited by iss (edited 04-08-2001).]

Nutty
04-07-2001, 03:26 PM
I'm a Dev Studio man myself! http://www.opengl.org/discussion_boards/ubb/smile.gif

cheers for the links.

Nutty

SUMECTAVAE
04-07-2001, 06:28 PM
Until man invents robots that walks decently, you'll never see a game where you can control all movements of a human body.

I'm sure ppl have tried to make games where you can move all your appendages, but how would they code in standing, walking, or balance? There're more muscles in the human body than there are keys on a keyboard, so some motor functions would have to be controlled by AI. Since there's no AI that can move with the ease of a human, the only option left for locomotion is dragging one's body across the ground.


It's simply too impractical to allow some1 to realistically control a human body without assisting AI, and since, as of now, AI isn't good enough to be of much help, it'll be a while b4 you see realistic games.


[This message has been edited by SUMECTAVAE (edited 04-07-2001).]

DFrey
04-07-2001, 07:47 PM
OT, but, Honda has in fact built at least 2 robots that walk upright on two legs. They can even navigate stairs.

SUMECTAVAE
04-07-2001, 09:38 PM
Yeah I know I've seen that. Do you walk like that? It walks like its crapped it's pants. I don't find it too impressive when ppl trump inventions that are supposed to mimic humans, and only mimic retarded humans. They really should keep that on the drawing board till its more fluid.

Munkay
04-09-2001, 03:09 PM
If I read it right he was only asking if rigid body dynamics have been implemented into a game to affect the human body and not if the entire body could be controled by the player.

From what I have seen Halo will be doing this using Inverse Knematics (eg. A Jeep hits a bump in the terrain, and bounces and the players body inside the Jeep jolts and shakes about in a realistic manner).

Also, I know Honda have made their robot look human but why should it walk exactly like a human? Our method of walking is actually very unstable (made up from a constant falling forwards in which we catch ourselves with each step), personnaly if I were to spend millions on an upright walking robot, I'd want to make sure it was stable and didn't fall over by giving it a different technique. Thats exactly what they have done, and I say well done to them!
How soon do you guys think it'll be when these physics will be implimented in games involving human bodies? Or have they already?

Munkay
04-09-2001, 03:15 PM
Opps! That last line wasn't meant to be there! http://www.opengl.org/discussion_boards/ubb/smile.gif

NeoTuri
04-10-2001, 05:03 AM
Anyone remember Trespasser? http://www.opengl.org/discussion_boards/ubb/smile.gif

A pure example of what too much control over the human body will do to a game.

Lars
04-10-2001, 06:34 AM
What about Black and White, u have a Creature there, and you can slab it very physically with your hand (mouse) http://www.opengl.org/discussion_boards/ubb/smile.gif
And even the the other physical interactions, for example with trees, or when throwing rocks are very realistic.

Lars

kaber0111
04-10-2001, 08:35 PM
Originally posted by Munkay:
If I read it right he was only asking if rigid body dynamics have been implemented into a game to affect the human body and not if the entire body could be controled by the player.

From what I have seen Halo will be doing this using Inverse Knematics (eg. A Jeep hits a bump in the terrain, and bounces and the players body inside the Jeep jolts and shakes about in a realistic manner).

Also, I know Honda have made their robot look human but why should it walk exactly like a human? Our method of walking is actually very unstable (made up from a constant falling forwards in which we catch ourselves with each step), personnaly if I were to spend millions on an upright walking robot, I'd want to make sure it was stable and didn't fall over by giving it a different technique. Thats exactly what they have done, and I say well done to them!
How soon do you guys think it'll be when these physics will be implimented in games involving human bodies? Or have they already?

i don't know about you..
but when i saw the video where they showed the guy drving that jeep...
it looked like it was a pain in the ass to drive.
the guy couldn't keep the damn thing steedy, he kept sliding, and when he kept screwing up in drving the jeep, he yelled out a shout like he was doing it on purpose.

and remeber, this guy has probably ridden that jeep a shit load.

personally i think halo will blow big.

anyways, back to rigid body physics..
there's some tutorials and stuff on the net..
but do your self a favor and _DONOT_ read them...
don't be bothered with the article in game programming gems either..
just read baraff's papers..
he starts into and goes on...
good luck.

[This message has been edited by kaber0111 (edited 04-10-2001).]

Nutty
04-10-2001, 11:58 PM
NOWAY!!!! Halo's gonna totally rock man!!

He was probably lame at driving it, cos he doesn't play games much.. maybe he was a manager or somat??

Looked like it had very weak gravity though, cos the jeep would stay in the air for ages..

I hope they sort out the lack of bullet marks on the environment too... I think it looks top.. can't wait to get my hands on it!

kaber0111
04-11-2001, 05:28 PM
well your entitled to your own opinion..
but that guy is the dude that made almost all the previous pr movies for halo.

and i'm pretty sure he's driven it quite a bit.



[This message has been edited by kaber0111 (edited 04-11-2001).]

EricK
04-12-2001, 04:29 AM
Originally posted by iss:
Rigid body dynamics is a large topic. Newcomers to computer graphics think that this is a new topic.

[stuff deleted]

One of the most interesting applications is combining Classical Rigid body Mechanics with Continuum Mechanics to do Deformable modeling. Demetri Terzopoulos www.cs.toronto.edu/~dt/ (http://www.cs.toronto.edu/~dt/) is one of the pioneers this topic.

[This message has been edited by iss (edited 04-08-2001).]

Sorry, but I just had to point out the irony here. If Dr. Terzopoulos were a pioneer in deformable modeling he would be about 200 years old! His body of work is very impressive but I do not believe he is a contemporary of Cauchy!