Page 102 - Human Inspired Dexterity in Robotic Manipulation

100 Human Inspired Dexterity in Robotic Manipulation

Fig. 6.10 Complicated bone shapes at the CMC joint of the thumb. Top: The common
mechanical analogy of the first metacarpal and trapezium bones. Bottom: The fixed joint
axes used for explaining different thumb movements. (Modified from Wikipedia, Trapezium
(bone)—wikipedia, the free encyclopedia, 2014, Available from: http://en.wikipedia.org/w/
index.php?title¼Trapezium_(bone)&oldid¼634623814 (Accessed 12 May 2015).)

When we model the kinematics of the human hand with the Denavit-
Hartenberg (DH) convention, we could treat the CMC joint as a simplified
saddle joint, and in turn, all the rotating axes can be nicely fixed (see
Fig. 6.10) and become suitable for the conventional mechanical design.
But in reality, the thumb motion could never be achieved with such a
mechanical substitute because the CMC joint requires not only saddle-
shaped surfaces but also a curved rotation axis that supports rotation, sliding,
translation, and pivoting motions [13]. In addition, the irregular shapes
of the articular surfaces are also responsible for stress transfer. It was estimated
that a tip pinch of 1 kg will generate 12 kg of joint compression at the CMC
joint. For a power grip, the load could become as high as 120 kg [14].
Previous studies on anthropomorphic robotic hand design were focused
on deriving the closest human hand model to build true-to-life robotic
hands. Many investigations have been conducted based on cadaver hands,
but only limited information was extracted and implemented into robotic
hand designs—in most cases just the number of fingers and length of the pha-
langes. We have not utilized the geometry of the bone structure, which con-
tains important kinematic information that could be used to greatly simplify
the design and prototype of anthropomorphic robotic hands. The biological
joints of the human hand have already been proven to work through years of
use in daily activities. It is of our interest to find a practical way to effectively
implement its functional features in robotic hand design.