I''ve been searching info on car/plane/etc physics but I can''t find anything suitable. Does anyone know a good website explaining such topics?

My most recent project has been vehicle physics modeling. I set out on the project more as a learning experience, and that is certainly been acomplished, I have learnt more doing this than any other project I've ever attempted. and the only reason I can give to this is that I never used an tutorials, examples, etc.. I only used my existing physics and maths knowlege.
had I simply copied and pasted from a tutorial, I would not have learnt anything. This is a golden rule too many people break.

Now I know intamatly how my program works, and if something very minor comes up, I know exactly why it's happening. I would not have gained this ability had I used a base tutorial, even had I studied it extensivly.

these, in my opinion, are what you need to know to do what I'm doing, there is no "car physics" in programming, it's all maths built on maths.

matrix maths, an absolute must.

basic plane/tirangle/sphere collision detection maths. another must. and, importantly, a complete understanding of how and WHY it works.

basic physics. forces and reactions, mass, centre of mass, etc. and especially how forces can be applied to a matrix for rotation. schools teach this (well, not the matrix bit).

Object attatchment, management, heirachy, etc.


and if you do find any information on emulating vehicle physics, you will have shot yourself in the foot before you begin. Simulate, not emulate. In my program, for example, there are objects for basic sphere collision detection. and more advanced extended objects that, say, rotate on one axis (wheels, steering), or move on one axis (suspension)... there is no car object. all these objects are extended from a base object that allows force application, centre of mass, child objects attatchment, etc.

I once did find such a document on vehicle physics emulation. It was all of 12mb zipped. and worked on emulating effects one by one. not working on what causes them.

even though I know it's not what you probably want, It will help you a lot more than what you were after. and you will learn.

Edited by - RipTorn on January 30, 2002 4:11:46 AM

Hi, take a look on my web site, i have a tutorial on how to build a linear flight model. www.web-discovery.net

glHorizon_Project



www.web-discovery.net

im looking for a funktion which gives me information about a intersectionpoint. horniestly the function should look like this:

vect3 findintersection( vect3 trianglepoints[3], vect3 linestart, vect3 lineend );

i need a triangle and the line to test. is the line goes throught the triangle, the function should return the intersectionpoint. if the returnvalue = lineend, the line did not vcontact the triangle. can anyone help me ? (i know, my english is godlike ))

im sorry... i am at work and i cant register cause of my firewall

i googled for that kind is raytesting. i could not find any site that offers that routine for free (only GNU). i search for a completly free routines!

thanks for help !

To anonymous poster :
What siaspete *roughly* says, in his 3 points, is :

1- first of all, browse the web yourself to look for indications. nobody here can believe such information is not already written clearly somewhere in the web.

2- you could register an account on gamedev.net''s homepage. it''s fast, it''s free, and it''s much better to identify ppl who post here. anonymous posts is really annoying. it should be used only in a last resort, eg you can''t login to gamedev due to cookie problems with your browser or things like that.

3- because your question is a bit off topic of this original thread, you should open a new thread. It will be clearer for the current thread (vehicle physics) and for your question (collision code). Moreover, you would benefit of a clear thread title. Actually, if ppl don''t care about physics but have the answer to your question, they won''t come into this thread because the title does not interest them.

Please, don''t think that we don''t like you or anything like that. We just try to make the forum as good as possible.
Thank you in advance for understanding

As for me, I gotta leave this place soon instead of ripping off this topic

Back on topic :
The complexity of the physics you develop have to take care of the "real-time ability" you look for your application.
If you want to develop agame, you won''t be able to use *true* physics : you''ll have to approximate a lot of equations.
If you want to create a simulation (for instance, to develop a graphics application for science research), then you''ll have to get as many equations as you can find, but don''t expect your application to run in real time (unless you optimize your code a lot, which is not what labs want).

That''s my 3c (eg 2c + half of 2c)

okay mr. anonymous poster.
I''ll open a thread just for you.
It''s hard to believe that the firewall blocks you, because I am behind a firewall, and that''s from here that I registered my account...