Tweet
Hello,
My name is Nathan, and I am a student doing researching arm
movements for mechanical assistors. I realize there have been many
previous discussions on properly differentiating and filtering data in
leg movements, but unfortunately I wasn't a subscriber to Biomch-L at
the time. I read several of the posts including those that refer to
David Winter, Graham Wood, as well as Pezzack, but after reading some of
these publications I find myself still somewhat confused. There seems to
be an inherent ambiguity in choosing a filter- meaning different filters
work better depending on the data and what you are looking for in this
data. This I understand. It seems however that splines and doubled or
reversing Butterworths are currently the most mainstream and widely
accepted forms of filters, with of course the added extrapolation
techniques (prediction and least squares for example) proposed by Giakas
and Baltzopoulos. Please let me know if I have misinterpreted anything
thus far.
With this being said, my questions then are these. If indeed
these are the most accepted forms of filtering, how does one go about
finding the best cut-off frequencies without having the actual
acceleration data to compare a RMSE to? Again if there are newer or
better filters please inform me of these as well. Also, in the paper
"Optimal Digital Filtering Requires a Different Cut-off Frequency
Strategy for the Determination of Higher Derivatives" by Giakas and
Baltzopoulos it refers to a procedure in which "displacement was
filtered with a different cut-off frequency depending upon optimal 0th,
1st, and 2nd derivatives." This procedure yielded stronger results than
others but again what are these cut-off frequencies and how can they be
determined without comparison data?
I don't mean to beat a dead horse by asking answered
questions, but any information would help. Please let me know if I'm on
the right track and additional information (i.e. papers, articles, etc.)
is always appreciated. Thank you all for your time.
Sincerely,
Nathan Manning
University of Washington
-----------------------------------------------------------------
To unsubscribe send SIGNOFF BIOMCH-L to LISTSERV@nic.surfnet.nl
For information and archives: http://isb.ri.ccf.org/biomch-l
Please consider posting your message to the Biomch-L Web-based
Discussion Forum: http://movement-analysis.com/biomch_l
-----------------------------------------------------------------
My name is Nathan, and I am a student doing researching arm
movements for mechanical assistors. I realize there have been many
previous discussions on properly differentiating and filtering data in
leg movements, but unfortunately I wasn't a subscriber to Biomch-L at
the time. I read several of the posts including those that refer to
David Winter, Graham Wood, as well as Pezzack, but after reading some of
these publications I find myself still somewhat confused. There seems to
be an inherent ambiguity in choosing a filter- meaning different filters
work better depending on the data and what you are looking for in this
data. This I understand. It seems however that splines and doubled or
reversing Butterworths are currently the most mainstream and widely
accepted forms of filters, with of course the added extrapolation
techniques (prediction and least squares for example) proposed by Giakas
and Baltzopoulos. Please let me know if I have misinterpreted anything
thus far.
With this being said, my questions then are these. If indeed
these are the most accepted forms of filtering, how does one go about
finding the best cut-off frequencies without having the actual
acceleration data to compare a RMSE to? Again if there are newer or
better filters please inform me of these as well. Also, in the paper
"Optimal Digital Filtering Requires a Different Cut-off Frequency
Strategy for the Determination of Higher Derivatives" by Giakas and
Baltzopoulos it refers to a procedure in which "displacement was
filtered with a different cut-off frequency depending upon optimal 0th,
1st, and 2nd derivatives." This procedure yielded stronger results than
others but again what are these cut-off frequencies and how can they be
determined without comparison data?
I don't mean to beat a dead horse by asking answered
questions, but any information would help. Please let me know if I'm on
the right track and additional information (i.e. papers, articles, etc.)
is always appreciated. Thank you all for your time.
Sincerely,
Nathan Manning
University of Washington
-----------------------------------------------------------------
To unsubscribe send SIGNOFF BIOMCH-L to LISTSERV@nic.surfnet.nl
For information and archives: http://isb.ri.ccf.org/biomch-l
Please consider posting your message to the Biomch-L Web-based
Discussion Forum: http://movement-analysis.com/biomch_l
-----------------------------------------------------------------