Make Marker a ModelComponent

[jimmyDunne]

[jenhicks]

Inherit from Station.

[aymanhab]

@msdemers Do you think Marker has to be attached to a Body (old implementation has a method getBody that's used ofr IK and visualization). Should RigidFrame provide that utility?

[msdemers]

Marker.getFrame() makes sense to me.
On Aug 29, 2014 4:44 PM, "Ayman Habib" notifications@github.com wrote:

@msdemers https://github.com/msdemers Do you think Marker has to be
attached to a Body (old implementation has a method getBody that's used ofr
IK and visualization). Should RigidFrame provide that utility?

—
Reply to this email directly or view it on GitHub
#46 (comment)
.

[aymanhab]

@msdemers True but low level utilities (e.g. InverseKinematicsSolver::setupGoals) do not know about intermediate Frames, all they know is how to track Markers on Bodies, and that's how they solve.

[msdemers]

Yeah, but that's just because RigidFrame didn't exist when the tools were
designed. There's no reason that these tools can't handle Markers fixed to
frames. In fact, they need to, because a modeling person is likely to
define a complex mechanism in a specific frame (maybe kneeFrame) and would
probably want meaningful markers to be derived relative to that meaningful
frame.

The tools will have to be augmented just a bit. A body and fixed frame are
both RigidFrames. Any RigidFrame can transform to any other RigidFrame, so
we have all the tools needed to make IK aware of frames instead of bodies.
On Aug 30, 2014 11:41 AM, "Ayman Habib" notifications@github.com wrote:

@msdemers https://github.com/msdemers True but low level utilities
(e.g. InverseKinematicsSolver::setupGoals) do not know about intermediate
Frames, all they know is how to track Markers on Bodies, and that's how
they solve.

—
Reply to this email directly or view it on GitHub
#46 (comment)
.

[aymanhab]

@aseth1 @msdemers @sherm1 @chrisdembia Since IK is a thin layer on top of the Simbody's method(s), having IK be aware of arbitrary frames requires a Simbody change or making frames (other than Bodies) into new massless bodies but that seems like pushing modeling complexity down to the computational layer. If a Marker is required to be based off a RigidFrame instead of arbitrary Frame than we can introduce a method RigidFrame::getBaseBody() to track any RigidFrame to the body it's affixed to without adding any complexity on the computational side. Thoughts?