[n/a]
The following is a list of replies to my original post, which asked
about the relative importance of passive limb mechanics to control, with
respect to gait prediction and synthesis.
I guess my original posting wasn't very clear, so here is a clarification.
My apologies to those who already saw this or similar stuff that I sent.
Most responses seemed to favor the "passive" system importance over the
"control" importance.
It happens that about 5 years ago, a fellow by the name of Tad McGeer
built some mechanisms, basically legs with knees and hip masses, that
walked down shallow slopes with NO CONTROL or actuation. Stability was
born from the passive interactions of gravity and inertia, and dissipative
collisions at kneestrike and heelstrike. This was
an extension of the "ballistic walking" concept of Mohon and McMahon.
In other words, walking is a "natural mode", if you will, of a set of
legs. However, parameter adjustments can change stability and efficiency of
the gait cycles, or can kill them altogether. Walking only occurs for
a relatively small subset of parameters. In our lab, the we study the effects
of these parameters on the mechanisms' gait dynamics and efficiency.
Now, the standard analogy goes like this: if you want to build a good
airplane, you put an engine on a good glider. If you want "good" walking,
you should probably start with a "good" set of leg parameters. Conversely,
if you are trying to predict gait patterns, if gait is "optimal" in any way,
it should be similar (by dynamics and control arguments) to the
"natural motion" of the legs; ANY
reasonable optimization would then produce pretty much the same gait.
Note that i am in no way claiming that walking is not controlled-
i am just saying that control is probably minimal, if you believe that
these models preserve the essence of human gait (which they may not, for
all i know). We haven't made them 3D yet, for one thing. Moreover,
legs dont just walk on level ground- they climb over things, and
jump and hop, etc.
But, it seems interesting that some parameter sets work and some dont.
Further discussion is welcome
-Mariano
Responses:
-------------------------------
Check out work by Ken Holt and his colleagues; use both medical and
psychology data bases. He's worked on precisely this question.
-----------------------------------------------
Wynne A. Lee, Ph.D.
Programs in Physical Therapy, and
The Institute for Neuroscience
Northwestern University Medical School
645 N. Michigan Avenue (Suite 1100)
Chicago IL 60611-2814
voice: (01) 312-908-6795
fax: (01) 312-908-0741
email: wlee@casbah.acns.nwu.edu
-----------------------------------------------
Mariano:
Your questions asked about gait analysis from a dynamic perspective, with
emphasis on passive (I assume versus "active") aspects. I am also assuming
that your use of the term "dynamical systems standpoint" reflects an
orientation toward a popular current perspective on control of
neuromuscular systems which is far from universally accepted, particularly
in traditional groups of "mechanical engineering" biomechanics. Thus my
responses are an attempt to describe a variety of perspectives within "the
biomechanics community". Note also that neuromuscular control is often a
multi-level concept, in which the "importance" of an idea may vary
depending on the level of analysis. Thus the importance of passive
dynamics might vary greatly if one looks at a level of a control system as
opposed to the level of action (muscle and bone).
>1) Has the passive dynamic concept been 'accepted' or 'rejected' by the
>biomechanics community?
Most, if not all, individuals engaging in research on the biomechanics of
gait include accurate representations of what I believe you are referring
to as the "passive dynamic concept". That is, they include both
theoretically and computationally the influence of mass and inertia of each
segment on all parts of the system. This accounts for dynamic effects
distributed among elements of the system based on both gravitational force
and the motion of each segment. Not to do so represents a fundamental
error in thinking and calculating; occasionally this occurs in a naive
attempt to use statics approaches to address dynamic problems. (I see this
in the work of some of my students.) Even those biomechanists who reject
(or are unaware of) the dynamical systems approach must account for passive
dynamic properties.
>2) What is the relative importance of control strategy vs passive mechanics?
My first response is "relative importance" to what? Relative importance to
a contoller, or to understanding observed movement, or to responses to
perturbations? I'll suppose you mean relative importance to understanding
normal gail patterns. If by "control strategy" you mean the actual
influence of the control system on the mechanical system, then the two are
of parallel importance. These two, almost by definition, comprise the
influences causing observed action. At any one instant, one or the other
may have a greater influence. For instance, to initiate motion in a still
and pendant leg one relies essentially entirely on "active" control
initiative. Of course the resulting action is affected by the passive
dynamical properties, but I think it is counterproductive to combine or
confuse control or causal influences with action or consequences. On the
other extreme, in mid swing the passive dynamic properties of the system
may fully account for the on-going action, albeit briefly, requiring no
"voluntary control".
I hope this helps. I assume you'll receive other responses, and I'd enjoy
seeing them and also continuing the discussion.
Larry Abraham, EdD
Kinesiology & Health Education
The University of Texas at Austin
Austin, TX 78712 USA
(512)471-1273 FAX (512)471-8914
--------------------------------------------------
Dear Mariano,
This a very interesting question you are raising.
>1) Has the passive dynamic concept been 'accepted' or 'rejected' by the
>biomechanics community?
Speaking strictly for myself, I say 'rejected'. Purely passive
movements are unstable because of the 'inverted pendulum'
character of human gait. This is why we have a central nervous
system. But passive mechanisms exist, and should be taken into
account.
>2) What is the relative importance of control strategy vs passive mechanics?
I would say that control is required, but that it is important
that the control input works together with the natural mechanics
of the system. For instance, natural frequences exist
(especially for the swing of the legs), and the control system
should not attempt to force the system to move at a completely
different frequency. This will lead to a large metabolic cost.
Could you post a summary of the responses to Biomch-L. maybe
including a count of votes for 'rejected' and 'accepted'?
-- Ton van den Bogert
Human Performancs Laboratory
University of Calgary
P.S. Don't forget to enter a Subject: line for Biomch-L postings.
-------------------------------------------------------------
Mariano,
Scott Delp and I have written a paper that addresses some of your
questions, at least with respect to the swing phase of gait. It is called
"The Influence of Muscles on Knee Flexion During the Swing Phase of Gait"
and will be appearing in J. Biomechanics in a few months. We used a
dynamic, muscle-actuated computer model to investigate determinants of knee
flexion during swing. We found that muscles do play important roles,
especially at the beginning and end of swing phase, such as the prevention
of excessive knee flexion by the knee extensors just after toe-off. If
you're interested, I could send you a pre-print of our paper.
Steve Piazza
____________________________________
Stephen Piazza
Graduate Student
Department of Mechanical Engineering
Northwestern University
piazza@nwu.edu
-------------------------------------------------------------
Mariano
I am very interested in these questions and have some opinions but no
hard data to back them up. One thing I've found strange is the lack of
discussion/study of the relationship between such mechanical issues and
function in prosthetics.
In prosthetics, the "natural"
mechanical properties of a limb, particularly a lower limb, are
presumably very important to the function of the limb. One thing we
often tell amputees undergoing gait training and/or
accomodation to a new prosthesis is: "Let the limb do the work for you."
In other words, the most efficient and normal-looking gait styles are
those in which active control is minimized. What we see with amputees
(and would be predicted from a purely mechanical analysis) is that a
natural and "low energy" gait style can usually be achieved in this
manner... but at one speed. To go faster or slower requires an increase
in active control. Also, to initiate a step (i.e. toe-off and swing)
requires active control by the amputee. I suspect the degree of this
energy input is greater for amputees than non-amputees control of the
ankle must be compensated for at more proximal joints which means that
the lower leg or the entire leg must be moved instead of, possibly, just
the foot.
In studying the questions you pose, a comparison between below knee
amputees, above knee amputees and non-amutees might lend a lot of insight.
Also, John Sidles at the University of Washington is doing some
modelling work which might be of interest to you in examining these
issues. His email address is: sidles@u.washington.edu
Sincerely,
Kim
Kim Coleman
Research Engineer
Prosthetics Research Study
720 Broadway
Seattle, WA 98122
Phone: (206) 328-3116
FAX: (206) 325-3607
e-mail: prs@u.washington.edu
Web Page: http://weber.u.washington.edu/~prs/
------------------------------------------------------
Hello,
>Since most of my literature is somewhat dated, I would like to know, or
>discuss with others, the following:
>1) Has the passive dynamic concept been 'accepted' or 'rejected' by the
>biomechanics community?
>2) What is the relative importance of control strategy vs passive mechanics?
Will it depends on which field you are interested in. If you are interested
in the behavior of the knee joint complex's components and the mechanics of
injury to the knee you would certainly stress "passive dynamics" and ignore the
control component. When a knee is injured, it's usually due to sudden load i.e.
there is no active input. Furthermore, the limiting case for knee components
is how they behave under sudden loading (with no active input agian) since you
would expect the active input to try to decrease the load on the knee.
On the other hand, if you are studying human gait, you have to take active
input and control into account.
There is a recent body of literature about dynamic knee models under the
assumptions of passive dynamics for two teams:
1- Mansour Moeinzadeh, Ali E. Engin and Tumer.
2- E. Abdel-Rahman and Mohamed S. Hefzy.
Eihab Abdel-Rahman, PhD
Biomechanics Labs.
Dept. of Mechanical Eng.
The Univ. of Toledo
-------------------------------------------------------------
Greetings fellow Cornellian,
I am a grad student of motor development and have read primarily that
literature. In that literature, the passive dynamics concept is thriving.
I have begun reading some of the robotics literature as well, but I am
still quite new to that area. I would be very interested in learning
about the responses you get from the list-serve. I hope that you can post
the results, or pass them along to me.
Thank you very much,
Leigh B.
Leigh F. Bacher
Cornell University
lfb2@cornell.edu
-----------------------------------------------------------------
Dear Mr. Garcia,
I read with interest your short discourse and questions of today about
control of the leg. I am assuming that you are talking about control of
the leg during the stance period of the gait cycle. I don't know that I
have any great answers for you, but I would like to ask you a few basic
questions to make sure that we can connect on the same wavelength.
Unfortunately (or fortunately depending on the way you view things),
podiatry teaches a lot about the kinesiology of the leg during the gait
cycle, but fails to utilize a lot of engineering terms. Therefore I hope
you will forgive me if I don't know everything about engineering. I am
interpreting your first question to be of the nature "does muscle control
play the dominant role in what the leg does or do the passive forces and
the inertial properties of the leg and the structures above and below it
determine how it acts in a gait situation." If this is not the question that
you are asking, please let me know.
As to "control strategy" question, I am assuming that you are referring
to the process of learned muscle control of the leg. Again if you are
referring to something else please let me know.
Now if I've correctly interpreted your questions, then I can give at least
some of the commonly held beliefs in the podiatry profession, which
we are presently teaching our students and practitioners.
We presently believe that Control of the leg during the stance period of
gait is more dependent in the transverse and frontal planes of the body
upon passive systems than upon any type of active system. One of
the reasons for making this statement is that it is almost impossible to
clinically change the transverse and frontal plane positions and motions
of the leg using any type of physical therapy or motor control therapy.
The major factors that seem to control rotation in the frontal and
transverse planes appears to motions in the foot and also the moments
placed on the leg by the superstructures above the leg. On the other
hand the sagittal plane rotations are more responsible to motor control
techniques and may be changed the most easily by motor control
therapies. It should be pointed out though that the inertial parameters
of the superstructure above have a tremendous influence on the
rotations in these planes.
Again your questions very much fascinated me, and I hope I have a
small understanding of what you are trying to ask. I would like to hear
more from you on this subject as control of the leg during gait is a major
therapy that we try to provide.
Sincerely,
Robert D. Phillips, D.P.M.
Professor of Podiatric Medicine
University of Osteopathic Medicine and Health Sciences
-------------------------------------------------------
It's a very interesting question. I'm afraid I haven't come across
any references on that topic, but I have recently been wondering about
that sort of thing. I'm not looking at gait, but at various knee
rehab exerices (eg isokinetics). I calculated the natural frequency of
oscillation of the lower leg in the sagittal plane (v. v. simple model -
just treating it as a rigid pendulum
with small oscillations) just out of interest, but I'm not sure whether
it helps me or not. But in any case, if you get any responses I would
be interested to see what people suggest.
Sorry I can be of any help.
Danielle.
$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$
Danielle Toutoungi,
Oxford Orthopaedic Engineering Centre,
University of Oxford.
tel: ++ (0)1865 227684
fax: ++ (0)1865 742348 email: danielle.toutoungi@eng.ox.ac.uk
"Time flies like an arrow, Fruit flies like a banana" - Anon.
$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$
--------------------------------------------------------------
Mariano,
You wrote ....
These interesting questions for which I do not have an answer.
The second question relies reasonably heavily on your own beliefs
on motor control programs etc. The most recent research efforts
initiated by Kelso about 10 years ago and refined by many of his
students and colleagues indicate that very little (if any) evidence
exists to support the notion of a motor program and that most
movements evolve based on the "passive dynamics" (i.e., inertia,
moment of inertia, gravity etc.). I hope you have some luck with this
topic and I would be interested in receiving a summary of your
replies.
Cheers,
Rob
---------------------------------------------------------
Hello. I recieved your email over the network and I have a couple
questions to ask. I am an American grad student over here in
Scotland studying knee kinematics and I am unfamiliar with what you
mean by passive mechanics-Is this in context with prosthetics or
rehabilitation? I would also be interested in receiving a summary of
the responses you get from your posting.
Thank you.
Spencer Malcolm
University of Edinburgh-
Mechanical Engineering
Want to know more? Email the CGA list! [n/a]
