Well, that's what the documents are for. Besides, you can watch the source code in C, matt! Read them, there's no way to avoid that. Besides, most of the public examples are poor documented, these programmers have been enlightened and they truly have transcended zen...
For example, start with "Simulating single rigid bodies". There, chapter 8, you'll find the code for using "RigidBody.c".
http://www.chaney.eclipse.co.uk/rigidbody.zip. Try to compile it. You can use that class in your program directly, it calculates quaternions and all, and forget about that part of your programming. Check the licensing with Mr. Chaney.
Then you can check the demo (
here, 750K) and the code source (
here). If you are a kart driver, you could learn a lot about weight displacement from the "extra" information of this simulator.
Richard Chaney 750 Kb car demo Notice the cool center-of-weight indicator (white big circle, top center) and the slip angle and tire slip data (to the left)
[img:512:384]
http://ciropabon.googlepages.com/Richar ... arDemo.jpg[/img]
Anyway, I'd say, start with a 2-D car. Don't worry about the representation, if the movement equations work there, they will work in the 3-D model.
The movement of a car depends on the forces on four patches "
the size of a human hand". All you need are the forces on the patches, no matter how you find them. The "main equation" is the relation between the size of these four forces and the position of the car (or its center of gravity) and its orientation, which is what RigidBody.c is for.
You can find the relation between steering wheel position and these forces on the patches the hard way, using a lot of magic numbers that are somehow related to the physical characteristics of the car.
You also can do it the easy way, taking, for example, the Pacejka magic numbers as a starter: they provide the lateral force you get when you move the steering wheel. They are pretty realistic; actually, there is no way to avoid using them that I know (well, not a practical one), which speaks volumes about how "empirically" we understand friction.
Braking and accelerating should be easy, you don't even have to relate them to any "physical" properties of the engine, as in an arcade game, where you can accelerate from 0 to 1 zillion kph in 3 seconds...

You won't go into clutch and differential influences at first, I'd say.
The dark art of the collision calculation (against grandstands or objects in the road, or other cars) can be left out for the second part of your project: don't worry if the car goes out of the road, simply "stuck it" for a couple of seconds, or reduce its "tyre friction", or simply, don't do anything in the beginning. Look at the simple solution you get in Car Demo: the car tilts 45 degrees when it reaches the edges...
BTW, please, keep us informed on how far you go, or if you can find a simpler alternative.
@Carlos: there have been quite a few threads on this subject; I learned a lot last year thanks to other forum members. Actually, if you haven't replied, I wouldn't have gone into writing anything else, I'm "mighty busy"...
