Biomedical Engineering Reference
In-Depth Information
6.2.1
Locomotion Principles
Robots presently display a wide variety of legged locomotion strategies (Hanson et al., 2003). Some
engineers explore sprawled-leg hexapod configurations akin to those of cockroaches, while others
build robots that hop on a singular leg. Others yet build robots that walk or run on two like a turkey,
a dinosaur, or even in the fashion of a human. And still others create robots that self-reconfigure to
adopt multiple locomotion styles (Fitch et al., 2000; Butler et al., 2002).
These robots are sometimes built just to resemble animals, while others utilize the physics of
animal locomotion and controls.
Unlike animals, most current legged robots do not employ mechanical compliance, instead
using rigid mechanical systems and, sometimes, computationally intensive digital PID loop con-
trols (although PID loops may be rendered much less expensive and intensive by using analog
elements). Animals, on the other hand, use a number of techniques involving mechanical compli-
ance to simplify controls and absorb shock, regardless of the number of legs. Full and Meijer (1999)
show that during animal-legged locomotion energy is stored in the spring-like compliant materials
as the animal lands in a stride, and then the energy is reemitted as the animal springs forward in the
stride, a dynamic that is akin to an inverted pendulum.
More than 20 years ago, Marc Raibert developed springy, dynamically stable-legged robots,
utilizing the principles described in the preceding paragraph (Raibert, 1983); numerous versions of
these robots bounced, ran, and turned flips to the world's delight (see Figure 6.1). Since then, many
robots have used this principle, as MIT's spring series elastic actuators, in Stanford's urethane-
rubber Sprawl series, and in many others.
In addition to absorbing energy, though, the viscous properties of the compliant material in leg
biomaterial also dampen the vibrations associated with locomotion (Full, 2000), thus stabilizing the
Figure 6.1 Early dynamic stability in action. (Image provided courtesy of Marc Raibert and The MIT Leg
Laboratory copyright 1990. With permission.)
Search WWH ::




Custom Search