The winner of this year's Ithaka Prize is Matthew Seabourne for his design and production of an animatronic hand.
Matthew explains how he developed his invention:
Aim: to create an
animatronic hand, using a system similar to that used by Special Effects
puppeteers for life like movement.
Find out more details about the project at http://goo.gl/k3tYvg
Matthew explains how he developed his invention:
Inspiration: I have been
interested in the special effects side of the film industry for a while,
especially what is called 'practical' effects (in other words, real, non CG[1]
effects). Practical effects are used when CG reaches its limitations and cannot
create something that looks real on screen. A lot of good directors also
believe it adds to the art of film making, to have to actually work with a real
piece on set instead of just adding the effect/object in post production.
The project is also a
perfect example of mechatronic engineering, which is the field I would like to
go into (so I can do animatronics as a job). The way the fingers work is purely
mechanical and they could be operated without servos, just using the fingers of
the operator to pull on the cables. The system used in this case to operate the
hand is an example of electronic engineering, using a Pololu Maestro 12 channel
microcontroller to operate 5 servos, one linked to each digit on the hand.
These two parts come together to form a Mechatronic system that can be operated
via computer, or a sequence of movements can be set and the system left to run
by itself.
The Results: As well as acting as
a useful talking piece featured on my personal statement, this project gave me
a greater understanding of working with electronics (especially servos) and was
a great deal of fun. It also resulted in generating (to date) a further 7700
views to my Instructables page, an invite to enter ABC’s Battle Bots TV show
and winning the PGS Ithaka Prize.
The making process:
Picking a design
The first mechanism I
looked at was created by FX specialist Rick Baker for the original Planet of
the Apes films (see the photo on the right).
These are operated by
the actor, who wears the prosthetic like a glove. It extends the actor's own
fingers, and moves as the actor moves their fingers. This works via the pulley
system that makes up the length of the rest of the finger. When the actor makes
a fist the string becomes taught and the fingers contract.
Prototypes
I reasoned that the
same motion could be achieved by using servos to pull the strings, so I built a
very simple prototype out of tin scraps and pulley wheels I found inside an old
robotic dog toy.
This replicated the
mechanism used in the gloves and upon testing proved I could use a servo to
pull the cables. I decided to make a neater prototype with the aim of then
building the final version.
Here however I
encountered a problem, I could get the finger to contract but could not get it
to relax to the original position again. Various experiments with elastic bands
and springs of various varieties proved ineffective and I reached a sticking
point.
After further
research I happened upon an 'instructable' for a 'simple animatronic' hand
(http://www.instructables.com/id/Simple-Animatronics-robotic-hand/). It was based
around cable tension and
plastic tubing. I built this as my next prototype, slightly adapting the design
to make it look a bit more human.
The hand is operated
by pulling the cables with each finger. Once the tension is released the spring
in the rubber tubing causes the fingers to straighten again. This provided a
useful second concept, the basic principle of which the final design is based
on.
Doing further
research I found the term 'cable puppet', these are the cable operated puppets
that are used in the special effects industry to control animatronics (although
sometimes they are operated by hand). I
found a university project that functioned exactly the way I was looking for my
system to operate (http://vegard.hammerseth.com/2009/03/school-project-mechanical-hand/395/. This is where most of the visual reference
for the final design came from, however I did modify a good 70% of it).
To understand the
mechanism fully (I had never worked with bike gear/brake wire before so had
little idea of how it behaved) I built a three fingered prototype out of PVC
pipe, small cabinet hinges, blutak and drinking straws (and gear cable):
The second prototype
(or MKII as I'll refer to it from now on) is worked in a similar way to MKI,
the operator pulls the cables and the corresponding digit flexes in a
human-like fashion. What is different is that it is not the plastic tubing or
the hinges that return the digit to it's original position, it is the bike gear
wire that naturally returns to a straight position.
The gear wire solved
the problem I faced when I first started out, the digits returning to their
original position, while also allowing for dexterity that was close to human,
meant I could move on and try to make a final (or complete, final suggests
there will be no further modifications to it in the future).
The Final Design
The final design was
made using copper piping for the fingers, cabinet hinges for the joints,
plumbing epoxy putty to fill in the digits, gear wire (I also experimented with
brake wire but found it to be too thick and inflexible for my purpose),
aluminium tubing and sheet and steel sheet for the palm. With the tools I had I
struggled to use solid joining methods (e.g. pop rivets) so after some trial
and error ended up using superglue to attach most of the parts together. This
probably presented the largest struggle as beforehand I had tried soldering,
drilling (to put screws through) and fixing with cable but none of these worked
or provided the strength needed to keep the digits from breaking apart while in
use.
I used servos with a
torque of 10kg/cm (a reasonably high torque) to pull the cables to operate the
fingers. To control the servos I used a Pollolu 12 channel Maestro controller,
which allowed me to plug in all 5 servos and control them separately via my
laptop. The Maestro software then allows sequences of servo movements to be set
and then played back in real time.
Evaluation:
The system delivers
what I intended it to do, act as an animatronic hand that could display simple
dexterity when it comes to digit movement. It is capable of grasping light
objects and should be able to pick them up and hold them with a little
adjustment to the cable-servo set up. If the hand had been used as a film prop,
multiple versions may have been built to perform different tasks for different
shots, or it may have been partially altered with CG effects later. The hand
has reasonably life-like movements, this may be improved in a later version by
using better servos (this was limited for this project due to cost issues) that
operate smoother and have a higher torque. I could also add bike tubing to the
inside of the aluminium rods that make up the forearm to prevent cable snag. Another
feature I could add to be able to control the servos 'live' (instead of
pre-setting a sequence) would be to add push buttons to move each finger.
If time and cost were
not issues, I would look at some way of articulating the wrist of the hand too,
potentially by putting small stepper motors or servos on the palm of the hand
to operate the fingers, leaving the wrist free of extra cables so that could be
operated without interfering with the digits. I would also look at creating a
latex or silicone skin for the hand (both of these materials are flexible and
can be used to imitate real skin) as this is what is done by VFX[1]
technicians to make the animatronic look like a live creature.
References: http://vegard.hammerseth.com/2009/03/school-project-mechanical-hand/395/
For visual reference for the design of the project
To fully document the
making of the final build I wrote an 'Instructable' (which can also be found
online here: http://www.instructables.com/id/How-To-Make-an-Animatronic-Hand/).
This lays out the full making process and also explains some of the challenges
I faced in more detail. It also shows the general public (hence why it is
written in a more informal style) that the project is something that can be
made at home with just a few tools.
[1] VFX-
Visual Effects
Comments
Post a Comment
Comments with names are more likely to be published.