PDA

View Full Version : Position Registration



Oddvar Arntzen
06-05-1996, 07:28 PM
------------------------------ Start of body part 1


Two weeks ago I posted the following message:

>I want somebody with user experience to tell me about advantages
>and disadvantages of the Flock of Birds and the Polhemus position
>registration system.

>We will be using such a system with 12 receivers to measure positions
>in a cube shaped space with each side 10 feet.

>I will also greatly appreciate tips about suitable acquisition and
>analyzis software.

>Answers will be compiled and sent to the list in about two weeks.

Here are the responses:

================================================== =========================

Hi Oddvar--

We too were looking into perhaps purchasing such a system, but alas, the
funding did not materialize. Nevertheless, I found quite a bit of
information on the two system (polhemus and flock of birds) in the biomch-l
archives. Below is the information from the "mini-FAQ" on how to search
the archives:

5. To browse or search the archives, use one of the following:
http://www.kin.ucalgary.ca/isb/biomch-l.html (preferred)
gopher://hearn.nic.surfnet.nl (browse only)
mailto:listserv@nic.surfnet.nl
GET BIOMCH-L LOGyymm (to receive an entire month)
INFO DATABASE (for help with keyword search)

Good luck on your search.

-Amy

Amy F. Polcyn, M.S. Biomedical Engineering
apolcyn@natick-amed02.army.mil
Geo-Centers, Inc.
US Army Natick Research, Development and Engineering Center
Natick, MA USA

================================================== =========================

There was an article on this in the last issue of Journal of Biomechanics


Jonathan Kofman, M.A.Sc., P.Eng.
University of Western Ontario
Engineering Science Building
Dept. of Mechanical Engineering
London, Ontario N6A 5B9
CANADA

================================================== =========================

I have used an older Polhemus unit. If data averaging is on it is an
accurate device; however, the sampling rate drops down to somewhere
around 20 Hz. This is too slow to capture many movements. Also, any
metal near the field causes havoc (after all, it is a magnetic digitizer).
Other than these points is a sturdy and effective tool.

Edward Lemaire
The Rehabilitation Centre
Ottawa, Ontario, Canada
613-737-7350 x5592

================================================== =========================

Antonio Valdevit at The Johns Hopkins University Biomechanics Lab has had
some experience and some success with The Flock of Birds and may be able to
help you. Please give him a call at (410) 550-6405. I am sorry but I do not
have his e-mail address handy. If I am able to locate it, I will provide you
with it as well. Hope this can lead to some help.

Steven A. Caruso, M.E.
Director, Biomechanical Engineering
Saint Vincents Hospital and Medical Center of New York
Phone: (212) 604-2504
Fax: (212) 604-2503

================================================== =========================

Oddvar,

We are currently using 2 Polhemus Fastraks to measure data from a
total of 8 sensors in a volume similar to that mentioned in your
posting. Each Fastrak can operate up to four sensors at a total combined
sampling rate of about 120 Hz (this means that if you are running two
sensors you are sampling at 60 Hz and with four sensors you are sampling
at 30 Hz). Multiple units can be synchronized to add extra sensors up to
a total of 16 sensors. Synchronizing multiple units does not effect your
sampling rate i.e. two units with four sensors each are still sampling at
a rate of 30 Hz/sensor.
We have tested both the standard Fastrak Source as well as the
Long Range Source. At this point I am in the middle of analyzing the
data, but I can give you a few comments and some rough estimates of the
error which we have seen.
We encountered two difficulties. 1) metal in the
test area can warp the magnetic field and cause errors in position and
orientation measurements - the original paper in J of B which described
the isotrak recommends a separation of twice the distance between the
source and sensor to the nearest large metallic objects. Beware, because
this includes metal reinforcement in your floor as well as ducting etc.
As you can imagine this can be quite prohibitive when working in a small
lab space. 2) as distance from the source increases so does the amount of
noise included in your measurements. As the distance increases past the
optimal operating radius (the optimal operational range is set at the
factory) there is also a small orientation based effect where angle and
position measurements are influenced by sensor orientation.
The first problem (metal distortion) is easy to solve, as you can
develop a calibration procedure to correct for these effects. We found
that small metal objects such as desks or chairs which were near the work
area had little effect on the data collected. However, our largest
errors were encountered when taking measurements near to the floor, which
was metal reinforced concrete.
The second problem is more difficult. Averaging or smoothing can
be used to clean up some of the noise. The orientation based effects
(which are small in comparison to the other two) are very difficult to
correct.
Finally, the choice between the Long Range and Standard Source.
We found that the Standard Source will provide accurate measurements up
to about 60cm. These measurements decay a small amount between 60 and 80
cm past this point the errors increase rapidly. At distance of about
1.5m the combined orientation/noise effects could lead to a 10cm/7 degree
variation in position and orientation measurements. At a distance of
about 250 cm the Long Range Source readings varied by about 4cm. At 450
cm this increased to about 35cm. Although the Long Range Source
effectively reduces the amount of noise encountered at long distances,
its results are less accurate at short distances because of its increased
sensitivity to fixed metal and a lower precision alignment of its
transmitter coils. I have not included any quantitative measures of
fixed metal distortion, as this should be approachable using a
calibration procedure, and it is going to depend on your lab conditions.

Hope this has helped you out.

Judd Day
Department of Mechanical Engineering
Queen's University
Kingston, ON, CANADA
3jsd@qucdn.queensu.ca

================================================== =========================

Hello!

I just got a package from "Ascention Technology" yesterday. They have a
chart comparing Birds vs. Polhemus.
Why don't you check on them through the web? Birds claim that they have
less distortion from metal field. But from my own experience, Birds was
interfered by unknown source on our campus. (A technical guy flew from
their company in east coast) Therefore, we decided to use
Fastrak for my thesis, studying head translation and
rotation.

I am also studying the differences between these two companies. So,
please let me know your result.

Alongkot Emasithi
University of Minnesota
emas0001@gold.tc.umn.edu

================================================== =========================

Hello Oddvar!

You were asking about "Suitable acquisition and analysis software"
Peak Inc. have a sytem for motion analysis, Peak Motus, it's a video based
system with acquisition-editing-calculation-reporting capabilities that well
meet the requerments stated in your mail.

Performance Technology in Stockholm, Sweden is the distributor of Peak Moyus
in Scandinavia and we would be pleased if we could do anything for you!

Best Regards!

Anders Norrman
Performance Technology
Karlslund
SWEDEN
perftech@bahnhof.se

================================================== =========================



Thanks to everyone who responded for their time and effort!

Oddvar.


------------------------------ Start of body part 2




Oddvar Arntzen
Faculty of Social Sciences
University of Trondheim
7055 Dragvoll
NORWAY
Phones 47-73591908 47-74827673
Fax 47-73591901
E-mail oddvar.arntzen@avh.unit.no


------------------------------ End of body part 2