Boy, was I confused! I'm looking at two Posts on the Forum and got them mixed up, and worried that I said something really silly because I was thinking about the other person's Post. But though I'm still confused, I don't think I grossly insulted anyone!
So this is about the Green Line and how to analyze it. Am I correct that we are looking at the X, Y, Z components registered by a Force Platform when someone who is standing on it jumps straight up in the air and lands back down on the platform making the very interesting multi-part "squiggle" (that's a Scientific Term that means "a lot of wiggly lines") shown in Green?
Before doing any sophisticated waveform analysis, I strongly recommend doing two things:
- Get a high-speed video at "ground level" of the platform, showing Takeoff and Landing clearly. You'd want to be able to go frame-by-frame and estimate the following time points:
- Beginning of crouch.
- Beginning of leg-straightening.
- Lift-off.
- Contact.
- End of knee bend.
- Resumption of upright posture.
- Acquire your Force data a little faster. Your DAQ Assistant seems to show you collect 100 points at 1 kHz (or 0.1 sec at a time), yet your plot seems to show 20 seconds of data, with the jump taking a second or three. There also seem to be some discontinuities in the data, possibly because your DAQ Assistant (which I totally dislike, and want you to stop using) seems to say you are taking finite (instead of Continuous) samples, and only 100 at a time.
First thing you should do is to find (use Google) "Learn 10 Functions in NI-DAQmx and Handle 80 Percent of your Data Acquisition Applications". Ignore Section 1, which discusses the Dreaded DAQ Assistant. Read the sections on how to use MAX to "play with" your acquisition device (what are you using, anyway?), and how to create a Task to do what you need (acquire 3 Analog Channels, differentially (do you know why? Your signal isn't too noisy, so single-ended might work)). Once you have your Task defined, you only need a DAQmx Start, DAQmx Read, and DAQmx Stop (plus saving the data, of course).
I could tell you about Producer/Consumer designs, which let you save as fast as you acquire, but why not take 5000 points at 2 kHz and tell your subject to jump when you say "Jump" (and say "Jump" right after you push the "Acquire" button). Now you should have a clearer picture of the components of Z force, and (especially if you have a linked video) can clearly see the phases of the Jump. Myself, I'd say the large negative peak is "leg extension" and the point it starts to rise sharply is when the foot leaves the platform (which bounces a bit, making the positive peak, then settling down to zero until BOOM, the jumper lands (but absorbs some of the shock in his/her knees, slowly straightening up).
Hmm. I just realized you cannot be using a Force Platform. When the subject is standing still, the force is essentially the weight of the subject, and when the subject is airborne, the force is zero. However, the "plateau" in your trace just before landing is at the same level as when the subject is standing ...
The point I'm trying to make is that you need to better understand your data before trying to interpret the little wiggles in it. Get the best (and clearest) data you can, try to confirm its time course, identify the segments of interest (and know what they are), then worry about analyzing it.
Bob Schor
