Hydraulics, sensors, and human modeling data-unified by proprietary
software-may help diagnose degenerative neuromuscular conditions
David Bak, Editor-in-Chief, Global Design News
Design News
May 7, 2001
The human body is a very unstable system of rigid components connected
by supple moving joints. When a man raises his arm, the observer notes
that particular motion. Not so apparent is the stabilization of all
the other joints in that man's body necessary to maintain balance.
The ability to measure and quantify these motions in real time may
contribute to the early diagnosis of degenerative neuromuscular conditions
like Parkinson's disease or multiple sclerosis.
That's because the stand-still movement of a healthy person falls
within a predictable range, or normalcy sphere. Tremors associated
with Parkinson's and similar conditions result in movements outside
this sphere. Not only are the deviations distinct and identifiable,
they increase in amplitude as the disease progresses.
Such is the rationale behind, as well as the observations resulting
from, a Computer Assisted Rehabilitation ENvironment known by its
acronym CAREN. Developed by Motek Motion Technology (Amsterdam,
The Netherlands; Manchester, NH) under a European Union funded program,
CAREN combines disparate technologies to help medical experts view
and analyze balance and coordination disorders as they happen in
an interactive, controlled environment.
"There is no standard network or communicative protocol for
treating balance disorders," states Motek CEO Edward Costello.
"Current evaluations are done visually. Because there are no
generic tools for diagnosis, there is little, if any, knowledge
transfer." CAREN, he claims, is the only system available to
register and evaluate human balance behavior in a repeatable and
controllable manner.
Combined technologies. Five parts make up CAREN: An SGI (Silicon
Graphics Inc.) Dual CPU Octane MXE computer, custom configured for
Motek; a flight-simulator-like motion platform; optical or magnetic
motion capture equipment; a human body model for simulation; and
enabling software that ties the different components together.
 |
| Electric and hydraulic actuators, responding to motion commands
from a central computer, control the motion platform on which
the subject stands. |
Called D-Flow, the patented software processes data from the motion
capture system, maps that information into the human body model,
and drives the motion platform. In addition, it performs all of
these functions faster than real time, defined by Motek as the visual
observation rate of 30 frames per second.
"To evaluate the subtleties of human balance behavior,"
explains Oshri Even-Zohar, Motek's vice president and chief technology
officer, "you need a system that is an order of magnitude faster
than the dynamics you are trying to measure."
Rexroth Hydraudyne B.V. (Boxtel, The Netherlands) supplies CAREN's
six-degree-of-freedom motion platform. Built to Motek specifications,
the hybrid system incorporates both hydraulic and electric actuators
to meet CAREN's required response times. During operation, the person
under evaluation stands on the platform which follows motion instructions
from the central computer while simultaneously generating new data
sets based on real-time biofeedback. It is this controllable and
interactive environment that gives doctors the means for analyzing
human motion.
A corresponding motion capture system, developed in The Netherlands
at the Technical University of Delft, employs infrared cameras,
passive optical markers, and a host computer to collect the subject's
exact movements. The retro-reflective markers, sewn into a body
suit or strategically attached to the subject's clothing, send IR
signals back to the cameras, positioned about the platform.
This information passes first to the host computer for processing
into position/orientation data, and then to CAREN's central computer
via high-speed Ethernet link. The proprietary D-Flow algorithm turns
the collected data into the degrees of freedom applied to Motek's
3D human body model—a stored library of common skeletal and
muscle motions that resides in the computer. "CAREN's ability
to reconstruct the moments of force around the joints of the subject,
and compare them against a standard database in real time,"
says Even-Zohar, "enables a doctor to view and measure a subject's
balance compensation strategies."
Faster rehabilitation. CEO Costello adds that
because CAREN identifies and quantifies how a subject stresses certain
muscle groups in response to platform motion, physical therapists
can use the system to devise appropriate treatment programs. Platform
movements can also be performed in close synchrony with projected
images to simulate every-day-life environments for aiding physical
therapy.

CAREN combines motion capture technology from the entertainment
industry with a flight-simulator-like motion platform for real-time
evaluation of human movement.
"The real-time feedback loop allows the operator to change
the environment as the subject is interacting with it," Costello
points out. Platform motion, he says, can be independent of the
subject's movements (as when standing on a very large boat), or
co-dependent with the subject's movements (as when standing in a
very small boat).
For example, a person with a broken leg naturally favors that leg
during rehabilitation, exerting less pressure than normal. By tilting
the platform slightly higher, a therapist can make the patient step
down sooner than anticipated. More pressure is applied to the leg,
shortening the rehabilitation process, and lowering treatment costs.
Even-Zohar envisions even more far-reaching medical benefits: retraining
people that have lost mobility. "We have a data motion capture
sequence of a healthy person walking," he explains. "Using
D-Flow, we convert data from the feet, knees, and hips into force
feedback instructions. Now, we situate the person who is immobile
over the motion platform with a harness, and attach their shoes
to the platform with Velcro. The platform's inverse kinematic motion,
according to the force feedback input, establishes a normal walking
pattern in the subject's limbs that may aid in retraining patients,
for example, after spinal injury.
Indeed, such experiments are presently underway at The University
of Groningen, The Nether- lands, where the CAREN system has been
integrated with other motion control and testing equipment into
a movement research laboratory. Run in cooperation with Beatrixnoord,
a Dutch rehabilitation center, the laboratory may provide new insight
into physiotherapy, orthopedics, neurology, and the early diagnosis
of a wide range of balance and coordination disorders.
CAREN is also being tested at the University of Amsterdam, and
the city's Central Orthopedic Technique.
Sidebar
Other applications
"Because the D-Flow algorithm is generic," says
Motek Chief Technology Officer Oshri Even-Zohar, "it
can be used in many ways outside the medical field."
For example, the CAREN system has direct application in the
simulation industry. It can be used to record and analyze
an airline pilot's motions during flight simulator training
to improve cockpit ergonomics. Another possibility: download
motion capture information into the black box of an aircraft
for reconstruction of a pilot's physical actions prior to
an accident.
Sports training is another potential area of use. Studies,
for instance, demonstrate that many athletes waste as much
as 30% of their energy when running by bouncing. D-Flow can
be used to reduce this tendency by capturing the athlete's
movements on the motion platform. The system intervenes to
relocate the platform surface where it eliminates the bounce.
"Not only does CAREN monitor progress in any motion
dependent process," Even-Zohar notes, "the system
lends itself to remote diagnosis via high-speed data link.
Sports therapists, doctors, and others can run full experiments
in real time from any distance, while remotely located workers
can simultaneously operate in an immersive force-feedback
environment as if they were in the same facility together." |
|