Help needed in prototype/tracking footwork and wrist motion

[nikosb]

I would like to create a device that tracks the position of the feet relative to the wrist. This is to be used outdoors for people that play sports and it is to be confined within an area of 15x30 meters. The idea is for specific times, not continuously, know the relative position of the wrist to the two feet within an accuracy of a couple cm. The horizontal position is very important, the vertical one less important. The orientation of the feet and wrist is also important to know. Any ideas how can that be accomplished? Would inertial measurement units work or other technology is needed?

 

[berkeman]

I would like to create a device that tracks the position of the feet relative to the wrist. This is to be used outdoors for people that play sports and it is to be confined within an area of 15x30 meters. The idea is for specific times, not continuously, know the relative position of the wrist to the two feet within an accuracy of a couple cm. The horizontal position is very important, the vertical one less important. The orientation of the feet and wrist is also important to know. Any ideas how can that be accomplished? Would inertial measurement units work or other technology is needed?

What ideas have you had so far? What constraints are there (can the mechanism be bulky, or does it need to be very non-bulky like if its used at a sporting event)? It certainly can be done with a mechanical apparatus, but that apparatus would get in the way a bit during, say, a tennis match or a basketball game.

 

[nikosb]

The idea I had so far is to use sensors on the back of the shoes like the ones that are sold for runners. Then there could be another sensor as a watch or wrist band around the wrist. That was my original plan. I am not sure which sensor would really do a good job. I was told that inertial measurement units have a lot of noise and error and it is had to extract relative position in an accurate and reliable way (within a couple of cms). Perhaps other kind of sensors would work better for this application but I am not familiar with what's available nor how well they would work. I open to suggestions

 

[Danger]

Could you use standard 3D motion-capture techniques and then set the computer to provide the relative positions?
I mean like the glowing dots on the body and high-speed high-def cameras or laser scanners.

 

[nikosb]

3D motion-capture technique is not suitable, I want something that doesn't rely on external input or help. I just started reading about sensors, my thought is to use ultrasonic trackers that do not need line of sight, coupled with inertial sensors that give the orientation. Knowing the relative position of the wrist to the both feet as well as the orientation of the feet and the wrist is all the information I need. Is that a possible solution?

 

[Danger]

Is that a possible solution?

I don't know, but probably somehow or other. Keep in mind that I'm not a scientist and in fact never graduated high-school. I just like mechanical stuff and trying to solve puzzles like yours. I can't see that Doppler-shifting of the signals as the sensors move relative to each other can possibly have any effect at the speeds and distances that you're dealing with, which I would wonder about for something really, really fast.
So, you'll need 2 sensors/transmitters on each foot (for orientation thereof), plus 1 or 2 on the wrist (depending upon if you have to record rotation as well as spatial position; let's call it the latter just to be safe). Each transmitter would have to record and output 5 different frequencies while ignoring its own, unless you can manage to multiplex the signals by time management. I know that the same applies to just about every possible approach; I'm just typing out loud here...
I'll go think on it some while doing other stuff and come back later.

edit: Have you tried approaching professional sports therapists, trainers, etc. for advice? They have some pretty incredible tech and some might be applicable to your project.

 

[Danger]

Hi Honey, I'm home...
Sorry, but it had to be said sometime, and when better than in a sports thread?
(That was just a joke, by the bye; I don't know to which sex you belong and I don't care.)
I was struck a few minutes ago by another thought, but I have no idea as to whether or not it will suit your needs. Although you have been reticent to divulge which sport your are studying, it would really simplify things for us to know. Anyhow, my extrapolation suggests that you are into Canadian or US football, based upon the measurements that you want (QB stance?) but the field dimensions are wrong for either. My very first thought was tennis, but there are those damned field dimensions again.
So... what I'm thinking of, if your permissions and budget allow, is to lay out an RFI antenna grid just under the grass and tag your physical points of interest. I don't really know anything about that technology, though, so it might not be appropriate.