Subject:
|
Re: Robotic simulators
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Wed, 27 Apr 2005 10:43:23 GMT
|
Viewed:
|
1194 times
|
| |
|
|
Hi Dan,
I'm sorry if I'm repeating anything anyone else has said in this thread but it
is not displaying correctly at all on my machine (this response seems to be its
own tree). Its making it very hard to actually take into account what else is
around as I have to search around and piece together order.
> Just so I understand... LDD (from what I can tell, it runs very buggy on my
> XP machine) is similar to MLcad with the added feature that you can
> test-drive movable parts (like a tecnic linkage -- not sure MLcad can or
> not). The problem with linkages would seem to be multiple solutions -- but
> if you start with a user-specified position, it seems like you should be
> able to compute it without ambiguity.
To be perfectly honest, I've used LDD very little (MLcad is much, much better)
and I'm not exactly sure what the problem they had was. It may be possible to
find the report on the web but I'm not sure where to start looking. I would have
thought that all that would be required for moving the hinges would be to
calculate a list of everything that moves and go from there with matrix
transformations but maybe even this becomes more hard when you try it.
> What you're saying was so difficult is collisions, so do you mean like
> actual parts moving around and bumping into each other?
Yeah, that's it exactly. The problem is that any numerical solution works best
when dealing with continuous forces/positions/momentums etc. When you take into
account hard collisions (like most materials in the real world), the continuity
is broken and the numerical solvers (like the ODE library that has been
mentioned in this thread) don't really work very well. I recently discovered
that there is in fact a whole field that studies what are called piecewise
continuous problems (like these) but I know nothing at all about it.
> I feel like I'm missing something. This is starting to sound like one of
> those things where you just don't get how messy it is until you try to
> actually code it. Yum!
I know the feeling. I recently decided to write a library to read in MLcad
files. Worked fine until I had to work out how to deal with MPD's... Fun!
Tim
|
|
Message is in Reply To:
 | | Re: Robotic simulators
|
| hi Tim -- --- Tim Gould <t.gould@yahoo.com>, (...) Just so I understand... LDD (from what I can tell, it runs very buggy on my XP machine) is similar to MLcad with the added feature that you can test-drive movable parts (like a tecnic linkage -- not (...) (19 years ago, 27-Apr-05, to lugnet.robotics)
|
2 Messages in This Thread:
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
