View Full Version : Re: Asimo robot gait

12-07-2007, 05:44 AM

As a Podiatrist I see this type of gait very regularly in the elderly and
those without the ability to build significant elastic energy in the
achilles tendon. This is typically known as an apropulsive gait, which is
fairly accurate in the elderly who have weak knee extensors but not a good
descritption of the younger people who can compensate by using the knee
extention power to augment the loss of spring from the Achilles tendon. The
faster these people run the more they flex their knees. I have a very good
friend who runs just like the Asimo robot and yet he runs for the local
running club and is not the slowest by a long chalk.

Usually these people have very / flat medial longitudinal arch and weak
posterior tibial muscle. This is often combined with an equinus ankle IE it
has resricted sagittal plane Range of Motion. Flexing the knee slackens the
Gastrocnemius and allows greater RoM thru the ankle. This is required to
enable uninhibited sagittal plane progression of the CoM. Unlike the usual
gait that is propulsive and has three pivots to progress forward, IE the
heel / ground interface, the ankle joint and the Metatrasal Phalangeal Joint
/ ground interface, this type of apropulsive gait has only two at best or
often one pivot to move thru the sagittal plane IE the ankle or talocrural
joint.If you look at the Asimo this is the case here. There is no heel /
ground or MPJ / ground rocker and he (it) relies soley on the robot
equivalent 'ankle joint' to move forward over the planted foot.

>From viewing the video from your link I would say the this robot sacrifices
efficiency of forward motion for stability.

What I mean by this is that the normal human walk is a progression of
falling forward and checking the fall with the next step. The CoM progresses
over the planted foot and the contralateral swingthru is delayed until the
ipsilateral foot is in the late propulsive stage of the stance phase of
gait. The body CoM is slightly forward of the foot and starts to fall toward
the ground after having been raised by the action of the Achlles and triceps
surae of the isilateral foot. The contralateral leg swings thru and GRF at
heel strike arrests the fall of the CoM and the three pivot action begins.
IE the potential energy become kinetic energy and keeps the CoM moving and
falling forward. This action is more exaggerated in running in the human.
The Asimo gait does not have very much of this vertical sine wave motion of
the CoM and the CoM or at least the main body stays close to the vertical
line of action thru the foot to the ground and not forward of it. Instead
the knee progresses past the foot while the contralateral leg drives the
robot forward. At the point of swing thru the knee is almost fully flexed
and ready to drive the next step. One would imagine that there are
relatively large moments about the knees during this action and for a human
to maintain this 'Grouch Marks' type gait would be physiolocically
inefficient and very tiring for the knee extensors. As Asimo has elctric
motors I guess this is not a problem but it probaly uses more electrical
power I should imagine. By following this gait style protocol and not
allowing the CoM to progress to far ahead of the planted foot the Asimo
reduces the risk of tripping and falling and so trades efficiency for
stability (in human terms at least). This type of maintainace of balance
can be see in martial arts and especially in Judo. Here the CoM is lowered
the knees are kept flexed the gait progresses in sliding steps so that the
body CoM does not progress far over or forward or outside the foot / feet
base of stance and does not go up and down in the classic sine wave motion
of normal walking. These strategies reduce the possibility of being
unbalanced and thrown. However the more this crouching style of posture is
adopted the less mobile one becomes and clearly Asimo is no ballerina but on
the other hand it may be quite difficult for him (it) to trip or fall
outside his base of gait.

It would also appear that Asimo does not have a sophisticated mechanism at
the foot for adaption to ground, which may mean he has to compensate thru
the knee and hip. Also the nature of the gait (as it appears to me and I may
be quite wrong) with its far more active leg propulsion rather than passive
falling from step to step creates a lot of torque that must be equalised by
counter torque from the upper body and arms. Perhaps this rotation counter
rotation also has a gyroscopic stabilising effect in the transverse plane.
Maybe this correlates to a torsion bar effect in the body that in effect
stores energy for the next step rather like the spinal engine model of
Gracovetsky that someoe else mentioned earlier I think. This may replace the
elastic (potential) energy storage of the achilles tendon and make stability
more achievable. Again this type of augmented propulsion is seen in walking
athletes who are not allowed to use the running action of bouncing off
elastic energy of the achilles tendon and instead use the exaggerated hip
action that loads the torsional energy for the next step.

So I guess Asimo is quite representative of a humanoid gait when that gait
has constraints and compensations imposed by physiological or anatomical
variations and or enviromental and performance criteria.

Hey this could be a great discussion Chris

All the best Dave Smith

FootHouse Folkestone UK
MSc Research Applied Biomechanics
Strathclyde University

----- Original Message -----
From: "Chris Kirtley"
Sent: Friday, December 07, 2007 12:18 PM
Subject: [BIOMCH-L] Asimo robot gait

> G'day all,
> This question is really just an excuse to be the first BIOMCH-L youtube
> reference :-)
> http://www.youtube.com/watch?v=cfaAiujrX_Y
> I saw the Asimo robot up close today here in Brisbane. It is impressive,
> but
> what interested me was the way it walks with flexed knees. When running
> the
> degree of knee flexion increases further. I am wondering if anyone knows
> why
> the Honda engineers did that - since the objective is to be as humanoid as
> possible, the flexed knee must clearly be such an advantage that they
> couldn't avoid it.
> Look forward to the discussion!
> Chris
> --
> Dr. Chris Kirtley MB ChB, PhD
> 608 Dockside
> 44 Ferry St.
> Kangaroo Point
> Queensland 4169
> Australia
> Tel. (07) 3891 6644 x 1608
> Fax 3891 6900
> Bond University
> Mondays (07) 5595 4430
> West End Family Medical Centre
> Wednesdays & Fridays (07) 3844 4111
> Clinical Gait Analysis: http://www.univie.ac.at/cga
> Book:
> http://www.amazon.co.uk/exec/obidos/ASIN/0443100098/203-6674734-4427132
> ---------------------------------------------------------------
> Information about BIOMCH-L: http://www.Biomch-L.org
> Archives: http://listserv.surfnet.nl/archives/Biomch-L.html
> ---------------------------------------------------------------