Re: [Algorithms] Algorithm to return point to surface of concave volume

[metanet s. <met...@ya...>]

> Since the capsules are mathematically defined by a length
> and a radius there are no surfaces so I suggest using
> standard collision detection instead. There are standard
> intersection tests that return a collision normal along with
> a intersection distance. I'm guessing you can find it in
> most open source physics libraries or just google it.

This is tricky, because the actual collision surface (i.e the outer surface of the union of capsules) is implicit; you can't simply apply a capsule-based method and expect a correct result. For example:

L1 S1
|  |
|  |
|_ac___S2
| *b
|__|___L2

L1 and L2 are capsule linesegments; the capsule surfaces (S1,S2) intersect at c. If you query at the asterisk, you should get c as closest point on the surface. There's no way to arrive at this through simple point-vs-capsule queries -- point-vs-L1 will give you b, point-vs-L2 will give you a.

You _can_ try to iteratively push the point out of each capsule and hope that it eventually reaches the surface; in the above case this would work, but in general it doesn't.

AFAIK there is no correct way to take the many local solutions you would get from individual point-vs-capsule queries and "munge" them together to arrive at a correct global solution. 

If there *is* such a method, I'd love to know about it though since it would be extremely useful. Currently we just use "iteratively project out of local collisions, and pray" :)

raigan



--- On Thu, 2/4/10, pontus birgersson <her...@ho...> wrote:

> From: pontus birgersson <her...@ho...>
> Subject: Re: [Algorithms] Algorithm to return point to surface of concave volume
> To: "GDAlgorithms" <gda...@li...>
> Received: Thursday, February 4, 2010, 10:36 AM
> 
> 
> 
> 
>  
> Andreas,
> 
> Since the capsules are mathematically defined by a length
> and a radius there are no surfaces so I suggest using
> standard collision detection instead. There are standard
> intersection tests that return a collision normal along with
> a intersection distance. I'm guessing you can find it in
> most open source physics libraries or just google it.
> 
> If you need increased performance you can always attempt to
> use proxy objects or perhaps switch primitive from capsule
> to something easier.
> 
> Hope that helps,
> Pontus
> 
> 
> Date: Wed, 3 Feb 2010 11:54:45 +0100
> From: and...@gm...
> To: gda...@li...
> Subject: [Algorithms] Algorithm to return point to surface
> of concave volume
> 
> Hi,
> 
> I have a point inside a concave volume, does anyone know of
> an algorithm to find the shortest vector from the point to
> the surface of the volume? The volume in this case is
> defined as the union of a number of overlapping capsules
> (lines with radius).
> 
> 
> Thanks 
> 
> Andreas Brinck
>  		 	   		  
> Skaffa en Windows phone så kan du
> chatta på Messenger var du än är! Skaffa en
> Windows phone så kan du chatta på Messenger var du än
> är! 
> 
> -----Inline Attachment Follows-----
> 
> ------------------------------------------------------------------------------
> The Planet: dedicated and managed hosting, cloud storage,
> colocation
> Stay online with enterprise data centers and the best
> network in the business
> Choose flexible plans and management services without
> long-term contracts
> Personal 24x7 support from experience hosting pros just a
> phone call away.
> http://p.sf.net/sfu/theplanet-com
> -----Inline Attachment Follows-----
> 
> _______________________________________________
> GDAlgorithms-list mailing list
> GDA...@li...
> https://lists.sourceforge.net/lists/listinfo/gdalgorithms-list
> Archives:
> http://sourceforge.net/mailarchive/forum.php?forum_name=gdalgorithms-list


      __________________________________________________________________
Ask a question on any topic and get answers from real people. Go to Yahoo! Answers and share what you know at http://ca.answers.yahoo.com