I am wanting to make an educational sport tool, which tracks the movement of body parts.
The aim of this is to let someone see when he is doing what wrong in his movement.
The Idea is that the person in question wears a harness with an arduino with three RFID scanners on his back and some RFID-tags on the joints of his body.
With scanning for the tags every 1/90 seconds that would make me able to get 30 FPS.
I would like the accuracy of it to be 5 cm.
This would mean that I have to be able to measure time to about one tenth of a microsecond.
Google has made me believe that one nanosecond might be the best I could get in assembly.
To have 30 cm accuracy I would have to have a precision of one microsecond.
Which is okay, but not preferable.
The arduino will store the time from when the tags responded.
Another computer will do the further processing.
I don't have a lot of experience in C programming, but I'll be able to get help from my school.
I never have done anything in assembly.
Programming help is not the main issue since my college teachers are very willing in helping people with hobby projects.
So my questions are:
What is the best time accuracy I can expect?
Will there be any problems regarding 12 or so RFID-tags close to each other regarding reading their name?
I'll be the first to admit that I don't know a great deal about RFID cards and readers, but my understanding is that the reader will detect a card when it is in close proximity and read (and write) any information stored on that card. When it is out of range (depending on the frequency they operate on, ~100mm) it cannot detect the card.
This seems completely at odds with the scheme you are proposing, which is why I asked.
dannable:
(depending on the frequency they operate on, ~100mm)
It's not that hard to find 50m RFID cards on the web.
They are actually fairly common as far as i can see.
So I think those will still work on a range of about one meter.
Over the range I'm not concerned, I'm more concerned about interference.
It's not that hard to find 50m RFID cards on the web.
They are actually fairly common as far as i can see.
Sure, but the readers are damned expensive.
Another issue with RFID readers is that they generally read a tag ONCE when it comes in range, and do not send any information when the card goes out of range. I don't think that RFID is going to suit your needs.
And, unless you got three different kinds of readers that read three different kinds of cards, you will have interference.
The basic idea, of using multiple radios to accurately estimate the position of a point in 3D, is not presently possible for amateurs.
However for many years, people having been tracking detailed movements of the human body. This is usually done by attaching LEDs to various points on the body and using images captured from video cameras to reconstruct the positions.
The other solution, but is for advanced users, is using image recognition, and you'd need a PC and a webcam or some camera, and programming oriented to images.