Concept: This above-elbow (transhumoral)
modular prosthesis is a design that can be used for rock climbing. Most rock
climbing prostheses don’t yet feature a hand that’s similar to a human hand -
that is, most of them end in hooks, only two fingers, or wedges that allow you
to grip with only 1-2 points of pressure. That's mostly because the technology for such a responsive and flexible artificial hand is still in the works. However, looking to the future, I’ve designed a functional piece that
would work like an actual hand, with articulating finger joints, as well as
grips on the fingers and palm that assist with traction during climbing. I drew
inspiration from tattoo designs around the world, primarily from an artist
named Kenji Alucky, whose work I’ve admired for years. His art is very
geometric, and has clean lines and stippling combined with negative space that
creates a visual feeling akin to sacred geometry. I also drew inspiration from
nature to create one more major visual element - a fern - to cover a portion of
the arm and complement the hard lines of the geometric piece.
Research: In my research on prosthetic devices for
above-elbow and below-elbow amputees, I continue to find that arms are much
more complex than legs in terms of musculature and joint function. There are
upwards of 30 muscles and at least 18 joints that work together to operate a
human arm, and the end purpose of all that complexity is the fine motor control
of the fingers (Richard F. ff. Weir, Ph.D.). One of my initial challenges was
designing joint articulations that are more flexible in range of motion than
your standard hinge joints. There are above-elbow prosthetic devices that use
only simple hinge joints, but you wouldn’t be able to rock climb very well with
that type of design. I took some artistic liberty with the overall engineering
of my device, and an engineer would be able to get me the rest of the way into
real-world functionality. My main goal was to create an attractive
rock-climbing cosmesis (outer coverage of a prosthesis) that could be fitted
with specific engineered parts later to become a functioning prosthesis, using
the engineering concepts I’ve proposed to fit inside it. For some
in-depth ideas on how to structure the joints, I found useful diagrams on the
Johns Hopkins Applied Physics Laboratory website. The lab provides a simple but
detailed visual aid on a sensored above-elbow prosthetic limb that they’ve
designed, with x-ray views that really assisted in planning out how I modeled
the fingers, elbow, and wrist joints.
Here’s what I’m working with in terms of the varying joint types:
elbow - hinge joint
wrist - combination rotating and hinge (ulna and radius
twisting)
thumb - ball and socket
palm - hinge halfway into the palm
origin of fingers - hinge in 2 directions at connection to
palm
finger middles and ends - hinge only
I also referred to Advanced Arm Dynamics for ideas on how to
design flexible joints for my model. Advanced Arm Dynamics is a rehabilitation
company that focuses on upper extremity prosthetics and rehab, as well as
providing access to new technologies to their patients. I referenced several of
their hand prosthesis technologies while modeling the joints for Scalar, but
the primary one is called the BeBionic hand. Its design allows for precision
gripping and flexible range of motion, which is what would be needed for a rock
climbing prosthesis. I would love to work with a company like this in the
future, since an advanced prosthesis like the BeBionic is simply begging for a
highly customizable shell. This company as well as Ottobock - the company the
BeBionic is now with - appears to specialize in myoelectric-controlled arm
prosthetics (which means the electrical impulses created by the person’s muscle
in their residual limb gets amplified and essentially powers the prosthesis).
Ottobock’s Dynamic Arm prosthesis was a design I referenced for creating the
elbow joint of Scalar. My design currently has a sort of pylon residing
underneath the shell, but I plan on modifying it to mesh more with the
myoelectric concept.
no naked edges
material assignments
hand detail
Modeling
PART 2 (from the halfway point):
Preparing to use FlowAlongSrf for the lower arm, I used
PictureFrame to bring in a reference of a fern tattoo similar to the idea I
wanted, made the object semi-transparent, then used that reference to draw
InterpolatePoints curves that would later become the fern cutout. I Rebuilt
these curves several times over so that when I used ExtrudeCrv (solid), the
surface would be clean enough to fillet the edges without getting a lot of tiny
naked edges that were unresolvable. After getting the curves simple enough, I
finally got around to setting up for my FlowAlongSrf command - I used
CreateUVCurves and chose the exact curves I wanted to reparametize from the
arm, and once I had the UV layout, I used PlanarSrf to get my first surface to
represent the lower arm with cutouts. I used Split to integrate the fern curves
with the planar surface, then moved the pieces I didn’t want to another layer.
I then extruded the whole planar surface, and my cutouts became three
dimensional. I used FilletEdge to smooth up the horizontal edges of the
cutouts. MatchSrf is a command I recently learned that adjusts the edge of a
surface to be tangent to the edge of another surface. I used this successfully
(once) when I was having issues with a naked edge in one of the fern surfaces.
It is not the magic bullet I was hoping for, but that’s par for the course. I
ended up going back and really sanitizing my curves before extruding, which is
always the most correct answer in the end. I was also having an issue with the
organic fern cutouts being a perfect closed surface before using FlowAlongSrf,
but afterward it would create bad geometry on the arm. The solution to that,
after several hours of trial and error, also turned out to be cleaning up my
curves from the very beginning, which I keep having to re-learn.
In drawing out the curves from the geometric concept art for the
bicep, I learned some new tricks when it came to defining the boundaries of 2D
shapes. I had never had reason to play with the circumscribed polygon tool
until now, and I found out there are multiple mathematical ways to start a
polygon and end it depending on its surroundings. It was just as useful as the
O-Snap options, because it meant I could draw out the imbedded
triangle designs based on the edges around them, which made the design look
really clean really quickly. I used Offset, Trim, and Mirror here a number of
times to ensure that the design was consistent in edge thickness, to maintain
cohesiveness and symmetry, and to manage overlapping curves before they caused
trouble later. I also learned how to use the Extend command during this process
to pull some curves to the extent of their boundary objects. When I tested
these curves on my surface using FlowAlongSrf, they were really skewed, so I
used Zebra to test the continuity of my surface. Getting some poor zebra
results meant I needed to rebuild my surface a little better, so I went back to
my original curves and built a new surface that would flow my pattern a little
more cleanly. After the initial test, I went ahead and started expanding the
detail of the bicep design, using Array to build the even spacing between the
offset bars. In testing the FlowAlongSrf some more, I found that the bicep
design was a little too busy, and didn’t complement the organic fern very well.
I simplified the overall design, and I’m much more pleased with the results. I
used Split to separate the geometric pieces from each other, but keep the
surfaces flush with each other. While I was using FilletEdge on the outside of
each shape, I periodically used Boolean2Objects just to check that I had
tangency but NOT overlapping objects. If objects are overlapping, a 3D print
will not work. Likewise, if you don’t have exact tangency, the pieces will not
print as a single object.
For the palm and finger grips/pads, I consulted with my rock
climbing experts on where to position the padding and how they should be shaped
in order to mimic the calluses that rock climbers build up on their fingers.
This provided me with accurate contact areas where it would be ideal to place
surfaces with increased traction. I alternately used ExtractIsocurve,
InterpolateCrvOnSrf, ExtrudeSrf, and OffsetSrf to get the curves and offset
surfaces I wanted to imitate from the original surface. I also edited some of
these control points to get a smoother shape, then used Pull to get them back
to the surface, because some points had escaped the surface somewhat. For the
palm pads, I Split the original palm surface with those surface curves, and
used JoinEdge with a duplicated surface to get the closed palm pad surfaces,
then moved them to another layer since they will be their own objects and have
their own material. For the finger pads, I created an asymmetrical eclipse
extrusion and kept it solid, then placed the extrusion through the fingers at
the correct angle for intersection. I then used Intersect, which I had never
tried before, to get the curves of intersection between the two objects. This
was a much better way than Project would’ve been to get consistent curves onto
angled surfaces. For cleaning up the original finger surfaces afterward, I had
kept copies of the small split surfaces so that I could use Untrim on the
original surface and it would be good as new. I used a similar process for
creating the angular cutout on the palm/hand surface.
While modeling the joints for the fingers and elbow, I relearned
the Orient command for positioning an object and aligning it with another
object. I vaguely remember being shown this command while doing a waffle
structure project a year ago, but I hadn’t really used it until now. It is
POWERFUL. I used this to insert and align bolts into the joint connections. The
joints themselves were modeled flat using curves, then using ExtrudeCrv
(solid). For the connections between the fingers and the palm, I needed to
create a way for the fingers to hinge forward and also move a little from side
to side. I used ExtractIsocurve from some of the hinges, then TweenCurves,
Loft, and BooleanDifference to create troughs in the palm surface for the
hinges to move back and forth in.
For the connections between the inner mechanism of the arm and
the exterior shells, I created simple lines then used Pipe so that I could
BooleanSplit them and keep the sections that would hold the two solids
together.
Materials:
This design has so many visible internal metal elements that I
needed to figure those out first before I could assign shell materials that
would complement the underlying structure. For the inner mechanisms of the arm,
I used a rough copper. For the screws and bolts, a blue zinc worked really well
to highlight the difference in material but not be too distracting. I then got
the colors and specularity figured out for the massive pieces, such as the
hand, fingers, and most of the arm. These I kept to a medium gray matte paint
for the fingers and upper arm. The hand and portions of the upper arm design
are a metallic gray paint. The details are a mix of an Axalta material called
Tempting Turquoise and a metallic lime green paint. The shell is a carbon fiber
material I had already figured out for a previous prosthetic accessory design.
The finger and palm padding/grips have a mold tech material applied to them,
which gives the impression of a slightly dimpled rubber material.
Sources:
DESIGN OF
ARTIFICIAL ARMS AND HANDS FOR PROSTHETIC APPLICATIONS: Richard F. ff. Weir,
Ph.D.
Amputee
Coalition:
Johns
Hopkins Applied Physics Lab:
Advanced Arm
Dynamics:
Ottobock:
Kenji Alucky:
