Tip:
Highlight text to annotate it
X
>> ALEXIS, JON (together): Welcome to a demonstration of our IK Solver!
JON: In this demonstration, we are going to use a four-jointed arm, all of the joints are ball joints.
As you can see, in this demonstration,
we have a path that goes through the very first joint
and we also have a path that tries to go
beyond the reach of the arm.
In response, our solver manages to get as close to the rest of the path as possible.
Notice the red arm does not act quite optimally.
That is because the CCD is a heuristic algorithm.
And, because of that, it is not optimum, but it is fast.
ALEXIS: Here's our lasso!
This is our arm in 3D, as you can see from the full range of motion,
and it follows a non-trivial goal that is within its operating space.
(Utter silence from the awesomeness)
ALEXIS: It's still going.
JON: Mhm. Our IK solver is pretty awesome.
ALEXIS: It's pretty fast: this is what amazes me.
JON: So how was your weekend?
ALEXIS: It was good.
JON: OK. Oh look, here's our third demonstration.
Oh wait, where's the circle?!
Wait...
ALEXIS: Hey, wait, wait.
JON: Oh, there it is!
ALEXIS: Ohh!
JON: OK. Oh, I remember! The circle was so far away from the arm that
we could not see it the first time unless we zoomed in.
And since it's so far away from the arm, the arm tries its best to get as close
to the arm (sic: goal) as possible, but its attempts are feeble.
Notice that it still traces out a circular path,
because in the CCD algorithm, it tries to minimize
the distance to the goal path, to the goal positions.
So that is why it still traces out, it still traces out
a circular path, yet the path isn't the one we want.
ALEXIS: I feel like we're torturing it.
JON: I know, right?! It just feels so sad.
Aww!
ALEXIS, JON (together): Thanks for watching!