Dear Renata,

This is a problem that many of us must have wrestled with at some
point, so I thought it worthy of a bit of thought. Here is my attempt to
enumerate the possible causes, in the order of the light path:

1. Infra-red strobe lights - they rarely fail, but note that the
capture volume needs to be within their power range. I presume this is
about 3-5 m long x 1-2 m high & wide for a Qualisys?
2. Quality of markers (they can get grubby and non-reflective with age
and wear). Some markers (e.g. sacral) are easier to see if they are on
a short wand.
3. Reflections - I had a lot of trouble once when a cleaner got into
the lab and polished the floor :-)
4. Camera lcoation - with four this could be quite tricky. I guess one
at each end quite high (1.5m or so), then probably two at 60 degrees
to the center of the walkway, but I'm not sure as I've always used 6
for bilateral recording. You will certainly not be able to get both
sides with four. Try to avoid one camera looking into the strobe of
another.
5. Camera lens - should be wide angle (2-35 mm focal length). Shorter
focal lengths can be used in cramped spaces but are not as good.
6. Camera stop (diaphragm) - not sure what this should be, but should
be checked.
7. Goes without saying that the lens should be focussed!
8. Camera angle should be adjusted by looking at the image seen on the
PC. Adjust so that there are minimal reflections and appropriate
coverage. All cameras do not need to see the whole walkway, but at
;east 2 (normally 3 in your case) should.
9. Thresholds - should be adjusted for optimum images in each camera.
10. Tracking software (AMASS) parameters - there are several and each
should be fine-tuned for best results (Qualisys willa dvise on the
usual defaults).
11. Extraneous lights - I used to turn some of the lab lights for best results.
12. Calibration - look at the SDs. They should be about +/-2 mm. If
much greater than this (say +/-5 mm - Qualisys will tell you what you
should expect) repeat the calibration. Make sure you sweep your wand
through the whole capture volume you intend to use. Not really a
tracking thing, but it's worth mentioning here that a Caltester type
routine should be done every now and then to confirm force platform
and kinematic alignment. See http://www.emgsrus.com/calteste.htm
13. Events (heel-strike, toe-off) need to be defined - in my opinion
this one of the most tedious stages! Although the software will
attempt to do this automatically, in my opinion you need to check each
one that is not defined directly by a force plate contact.
14. Marker placement - thigh wands in the Helen Hayes model are
particularly tricky to get right. They need to line up with the hip
and knee joint centers. See http://www.univie.ac.at/cga/faq/reliability
Knee marker is easy to place too anteriorly or
posteriorly. See http://www.univie.ac.at/cga/faq/hayes
15. Static trial - do you do one?
16. Labeling - are the right and left legs getting swapped - all labels correct?
17. Any marker trajectories crossing?
18. Anthropometry correct?

If I've missed any perhaps someone will remind me. I realise it's a
long list, but each of these factors can degrade your system's
performance. It can take a while to get it all optimized!

Chris
--
Dr. Chris Kirtley MD PhD
Associate Professor
Dept. of Biomedical Engineering
Catholic University of America
Washington DC 20064

Gait: http://www.univie.ac.at/cga
Radio: http://radiolistener.blogspot.com