CGA FAQ: Segment Powers
from CGA 8/96
Date: Fri, 9 Aug 1996 14:01:34 +0800 (WST)
Reply-To: [n/a]
Originator: cga@info
Sender: [n/a]
Precedence: bulk
From: "Sang-Hyun Cho MD."
To: Multiple recipients of list <[n/a]>
Subject: How can I use AMTI force plate for small children ?
MIME-Version: 1.0
Hello, this is Dr. Sanghyun Cho at Seoul, S-Korea.
We are using the Vicon 370 gait analysis system with AMTI and Kistler
force plates. And we have not been able to do kinetic analysis for
small children because they usually touch AMTI force plate more than 3
times per trial due to their short step length. Our AMTI force plate
is 51 cm long and 46 cm wide, and we placed two AMTI forceplats serialy
adjacent.
Can anybody help me with an idea that can resolve the above problem ?
Sanghyun Cho MD.
sanghyun@bora.dacom.co.kr
Yonsei University, College of Medicine, Rehabilitation Hospital
SEOUL, S-KOREA.
Dear Dr. Cho,
Ha! Someone else hs the same problem! You may have noticed that I listed this on the CGA Frequently Asked Questions, and was hoping someone would bring it up.
We also have an AMTI plate (61 x 61 cm) installed here for sports research, so it's ridiculously large for kids. As you'll see if you look closely at the videos, we try to get the child to walk with one foot on the plate and one on the floor by the side, so they walk at the rear of the plate right to left, then at the front left to right. Even so, of course, it's still difficult but you should note that you can still use the recording if there are two successive falls of the _same_ foot, since there will be a break in the recording to separate each foot-fall. The key thing is to get the child to walk with the foot you're not interested in off the plate.
This can result in a lot of tedious repetition, during which time the child (and more especially!) the parents become frustrated. I often wonder about the validity of recording a stride under these circumstances!
This is one reason why I brought up the topic of segment powers, since it seems to me that if we could get something from them, we could avoid the force plate altogether. I wonder whether the improved validity in terms of recording a more natural stride would compensate for any loss of information from the force plate.
It's nice to see such an issue (which is never mentioned in the journals) getting an airing in public at last!
Chris
Date: Sat, 10 Aug 1996 03:55:34 +0800 (WST)
Reply-To: [n/a]
Originator: cga@info
Sender: [n/a]
Precedence: bulk
From: "Andrew W. (Drew) Smith"
To: Multiple recipients of list <[n/a]>
Subject: RE: How can I use AMTI force plate for small children ?
MIME-Version: 1.0
I don't mean to muddy the waters, but it seems to me I recall a paper from the
Winter group (I think Gord Robertson was the main author) which showed that
it is possible to determine powers from kinematic data for most segments quite
accurately (except for the foot segment). It involved determining the segmental
energies and then integrating those.
I hope I don't get blasted for suggesting this approach - I've never explored it
to any extent and I don't have the paper in front of me at the moment. It's just
something to consider.
Cheers,
Drew
__________________________________________________________________
Andrew W. (Drew) Smith, PhD
Director of Research Assistant Professor
Lyndhurst Hospital University of Toronto
Toronto Ontario Canada
awsmith@cycor.ca +1 (416) 422-5551 x3040 FAX: +1 (416) 422-4380
__________________________________________________________________
From: HowieDBPG@aol.com
Date: Fri, 9 Aug 1996 08:24:21 -0400
To: ikirtley@info.curtin.edu.au
Subject: Segmental Power
Chris,
Are you familiar with the work of Tad McGeer? In 1993, he published a paper
entitled Dynamics and Control of Bipedal Locomotion in the Journal of Theor.
Biology (163, 277-314). In it, he describes the concept of passive dynamic
walking. I have been quite fascinated by this because it suggests that
humans are designed to walk without very much in the way of muscular input.
Although muscle activity is required for starting, stopping and pertubation
(sp), walking occurs rather effortlessly. If he is correct, then we may be
looking at gait from the wrong parameters. When muscle action is greatest,
we may be looking at the most "inefficient" action. I believe that this
would be a fascinating discussion for the group. I look forward to your
comments.
Regards,
Howard J. Dananberg, DPM
Dear all,
I thought I'd answer all these three comments on my post about segmental powers last week in one reply...
I got a question for you. Are you reading "The Biomechaincs and Motor
Control og Human Gait" by Davis Winter ? My copy is the second edition
and I can find only data tables on p.133. If you mentioned other book,
could you give me informations about the book with your opinion about it
? Publisher and ISBN will be great help for me.
---***---
Just a short comment regarding your statement "This is one reason why I brought
up the topic of segment powers, since it seems to me that if we could get
something from them, we could avoid the force plate altogether."
Because power is computed from:
Power = Torque x omega
and the only reliable way to compute torque at lower limb joints is through
inverse dynamic analysis, it seems to me that one could not compute joint power
without the information the force plate provides. Of course forward dynamic
analysis could be used to resolve the torques at the joints without the force
plate, but the accuracy of upper body kinetics is very suspect (relative to the
accuracy of a force plate) due to the difficulties in accurately the computing
trunk center of mass and moment of inertia - I think we are stuck with the force
plate problem.
---***---
I don't mean to muddy the waters, but it seems to me I recall a paper from the
Winter group (I think Gord Robertson was the main author) which showed that
it is possible to determine powers from kinematic data for most segments quite
accurately (except for the foot segment). It involved determining the segmental
energies and then integrating those.
------------ reply -----------------------------------------------------------
The book I referred to is indeed Biomechanics & Motor Control of Human Movement by David Winter, but I'm afraid the section I mentioned is only in the Second Edition (1990) John Wiley ISBN 0-471-50908-6: Section 5.5.2 Power Balance within Segments, pp 133-136.
The comment about Joint Power is correct, but note that I was specifically wanting to talk about SEGMENT power, which is calculated by differentiating the sum of potential + translational + rotational energy:
E/t = mgh + (mv^2)/2 + (Iw^2)/2
Since all these quantities can be derived from kinmeatic data alone, I was wondering if this might be one solution to avoiding the problems of force plates in young children. I am not saying that segment power is the same as joint power, only that it might somehow provide some of the information that we normally get from the joint power.
Oh, and by the way, my understanding of forward dynamics is that it is a method for predicting the kinematics based on knowledge of the muscle activations by such mathematical techniques as Lagrange. I think in the approacg of working out joint moments and powers from head down rather than foot up is still an inverse dynamics analysis. Maybe I'm mistaken here?
Now, as for that paper by Winter, I know exactly the one you mean, but unfortunately lent my copy to someone. We became quite interested in this method for a while, but I recall that someone informed me that a subsequent artcle by Alishensky suggested flaws in the method. I confess I've never really understood what these were, so I'd be grateful if someone could tell us.
The other nice thing about the power balance, by the way, is that it provides a means with which to double-check the accuracy of your inverse dynamics. The sum of:
Fdx.Vdx + Fdy.Vdy + Md.ws + Fpx.Vpx + Fpy.Vpy + Mp.ws
should equal the power as calculated from the kinmeatics (above equation).
where Fdx = joint force at distal joint of segment s in x-direction etc. and Md.ws = Moment at distal joint x angular velocity of segment.
I've just started to run this kind of check on my own inverse dynamics, and I must be honest and say that so far I can't get it to balance! So either my inverse dynamics or this equation is flawed. (I hope it's the latter!).
Anyone else do power balance calculations, or using segment power in any novel way?
Chris
Want to know more? Email the CGA list! 
[n/a]
Back to Clinical Gait Analysis home page