Archived

This topic is now archived and is closed to further replies.

_walrus

modeling more complex mass representations

Recommended Posts

Generally most articles and discussion i''ve come across dealing with physics is on simple point mass representation, where the object''s physics are handled at one point in space (usually the center of the object). I''m just wondering if anyone knows of any articles or tutorials that deal with modeling a more complex physics system. For example objects that have masses that are not uniformily distrubuted: suppose i shoot an arrow into the air straigt up, eventually the arrow head will be the first part to hit the ground since it is heavyest at the head. The trivial solution would be to just rotate the arrow about some axis that is perpindicular to the gravity at a point near the top of the arrow as it is desending, but I want to be able to apply these physics to other shapes, not just arrows. Thanks

Share this post


Link to post
Share on other sites
As well as the position and quantity of mass the other main factors in such a situation are the moment of inertia and in particular the aerodynamics. Moment of inertia is pretty straightforward: look at any decent dynamics reference.

For an arrow aerodynamics is the most significant, and unfortunately it''s one of the most difficult topics in physics: most games either ignore it or fudge it. It should be possible to fudge it for a flighted arrow as it is designed to have a particular, relatively simple, aerodynamic model. But such a simple model will not appy to a complex more general shape, and generalised aerodynamics is not a solvable problem on today''s game platforms.

Share this post


Link to post
Share on other sites
I'm no expert but an inertia tensor will give more appropriate distributed mass approximations than a single point mass, also something interesting might be to use a set of point masses within the object linked by rods to explicitly simulate the mass distribution (a realtime approximation of course). Of course as the previous poster said there is much more to fudge in a simulation than just mass distribution.

[edited by - JuNC on June 9, 2002 10:11:42 AM]

Share this post


Link to post
Share on other sites