by John Nagle <nagle@[EMAIL PROTECTED]
>
Jun 25, 2007 at 07:25 PM
bpj1138@[EMAIL PROTECTED]
wrote:
> On Jun 2, 4:19 pm, utkualtink...@[EMAIL PROTECTED]
wrote:
>
>>Hi,
>>
>>My impulses are not enough to stop moving objects, they slowly
>>interpenetrate while there are continuous forces effecting them. This
>>fact is pretty well discussed in this
thread:http://groups.google.com.tr/group/comp.games.development.programming....
>>
>>I have changed my imlementation to move objects out of penetration
>>using just their LINEAR velocity. It works however the objects are not
>>stable and tremble. I can maybe solve this using collecting impulses
>>in more than one pass and apply later but before doing that I want to
>>be sure I am doing right by using just linear velocity, I have tried
>>to move only contact points and simply added an integration
>>process(linear and angular) using final linear velocities of contact
>>points, but it makes the stuation worse when micro collisions are
>>happening... So Is it right to use linear only solution?
>
>
> So what if the object is resting on an edge of a table, for example,
> and most of the object is over the edge? In your system it will never
> be rotated and fall off the edge. I'm just sayin'..
>
> --Bart
>
Trying to do resting contact with impulses is never going to work
all that well. Read Brian Mirtich's PhD thesis.
http://www.kuffner.org/james/software/dynamics/mirtich/index.html
John Nagle
Animats