Chronology Current Month Current Thread Current Date
[Year List] [Month List (current year)] [Date Index] [Thread Index] [Thread Prev] [Thread Next] [Date Prev] [Date Next]

Re: [Phys-L] Simone Biles Triple Double in Slow Motion

Let me take things out of order:
On 6/12/21 9:32 AM, Dan Beeker via Phys-l wrote:

The one caveat I have is motion
analysis generally takes considerable time and some experience. Start
with simple things and work up to more difficult motions. Such
analysis is a good skill builder but best learned incrementally.

We agree on that.


5. Motion analysis from videos require some skill in making
assumptions. E.g. where is Simone's c.o.m.? For a typical person
their center of mass would be approximately at their belly button.
But Simone is really muscular so may be atypical. A good question
would be to ask "Can one verify Simone's center of mass using
information from the video?" Perhaps this is bootstrapping? Again
another lead in to better understanding of how to solve problems by
making and verifying assumptions.

That makes several bad assumptions.
The CM is not "attached" to the body anywhere, typical or otherwise.
When she folds her body into a V shape, the CM is well outside
her body.

High-jumpers and pole-vaulters exploit this fact. Their CM passes
/under/ the bar even though every part of their body (sooner or
later) goes over.

1. An object in the earth's gravitational field falls with an
acceleration of 9.8 m/s each second

True of the CM, neglecting aerodynamics.
Not true of a black box with internal moving parts.

For example, consider the free inertial motion of an opaque
boxcar with excellent bearings. It contains a bunch of clowns.
They all run to the front, then to the back, and so forth. An
external observer will find the motion of "the" boxcar as a
whole very complicated and impossible to predict.

It's even worse if it's a spacecraft or the like, free to move
in 3D and rotate around 3 axes.

A gymnast is not quiiiite that bad, insofar as you get to observe
most of the moving parts, but you have to put in a bunch of work
to observe them closely.

2. An airborne object's horizontal velocity (assuming horizontal
velocity is slow enough for drag to be negligible) doesn't change
doesn't change significantly.

Again, not true of a black box with internal moving parts.
Unless you do really complicated motion capture, you will
not know where the CM is. Not even a usable approximation.

4. Getting good results requires thoughtful placement of the camera.
Not so thoughtful placements can lead into discussions of why 1, 2
and/or 3 do not appear to be valid.

Decent motion capture requires multiple cameras and/or accelerometers.
Even then it is hard to get good data. Then it takes a whole bunch
of computational physics to make sense of the data. That includes
exploiting the various known constraints (what's connected to what).
Serious people with serious budgets have worked on this.

There's a big overlap with the robotics literature.

Although there are better ways to analyze human motions camera views
are not really so terrible to analyze. There is free motion tracker
software. I believe Vernier and Pasco , and perhaps others provide
analysis software in their offerings.

I guarantee you that their software is not good enough to analyze
anything as complicated as gymnastic motion. Not even close.