Archived

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

Collision detection with angular momentum

This topic is 6033 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

Recommended Posts

These are 2 separate topics really but I thought I would just throw them together since they do connect in some form at the end. I''ll start with my questions on collision detection. I have several oddshaped polyhedras. These include anything from animated characters, to simple geometric objects, to odd shaped things you would find on a desk. My question is what would be the best method of collision detection? I''ve been doing alot of reading on the subject for quite some time now and am pretty stumped. I''ve determined that due to the nature of the environment it would be best for me to use a broad to narrow phase approach to this. However for both approaches, i''m fairly clueless as to which method is best. When I say method, I mean such as using AABBs, Bounding spheres, OOBs, etc. Any advice on this subject and some whitepapers would be greatly appreciated. I''ve read the previous forums about using SWIFT and have been looking into that as well. Ok, my next topic. Angular momentum. This plays a very strong role when simulating objects in games. I''m having some trouble understanding angular momentum however. Correct me if i''m wrong but I see a strong similarity when comparing it to linear momentum. Similar equations and the such. Its all still very new to me and any articles or simple papers on angular momentum would be greatly appreciated . Now how these link together is pretty straight forward. When an object is spinning at a certain speed, it hits an object with a certain force and depending on the physics properties of both objects, reacts in a certain way (bounces off, breaks the object ect.) When reading up on collision detection methods, not many mention this sort of movement in the objects and most tend to deal with linear movement. So to sum things up, i''m looking for a good method of collision detection that takes into account both linear and angular momentum as well as some more simple papers on angular momentum. Thanks a bunch.

Share on other sites
My first suggestion is that you do a bit of coding and experimentation. Reading is nice, but you''ll find it easier to answer your own questions if you experiment with techniques you''ve already read about.

In two dimensions, it is straightforward to deal with rotational effects and angular momentum. Reason being the equations for linear momentum conservation are independent of the equations for angular momentum conservation. But when you go to three dimensions, the equations are coupled. And the coupling is what leads to gyroscopic effects, which are a bit tricky to simulate. For example, this is one of the cases where the old crappy simple Euler integrator breaks down and fails to work nicely.

At your level of understanding, I''d suggest you look at the book "Physics for Game Developers" (David Bourg). It''s an introductory type book. It does meander into areas that aren''t really important, but it should sufficiently cover collision detection and angular momentum for simple rigid body physics.

Alternatively, look at Chris Hecker''s personal pages at his company''s web site, www.d6.com. There''s a series of (I think) 4 introductory game physics articles in PDF format, which were originally published a few years ago in Game Developer Magazine. As I recall, he does go in some detail on the 3D inertia tensor, which is how the linear and angular equations become coupled. His articles are fast paced, and perhaps a bit harder to follow than Bourg''s book.

Also, the gamasutra site www.gamasutra.com has article archives from Game Developer Magazine. Its free to register for the site. There are numerous articles there on practical techniques for collision detection and physics including rotational effects.

You might want to search the gamedev forum archives as well (see the cool search button at the top), and look at the articles and resources section.

Graham Rhodes
Senior Scientist
Applied Research Associates, Inc.

1. 1
Rutin
46
2. 2
3. 3
4. 4
5. 5
JoeJ
19

• 11
• 15
• 9
• 10
• 13
• Forum Statistics

• Total Topics
633004
• Total Posts
3009836
• Who's Online (See full list)

There are no registered users currently online

×