|
Large-Scale Sensing
A living organism like the human being contains
millions of microscopic sensors scattered throughout
the living tissue. Human skin contains millions
of fully integrated sensors, including those for
cold, heat and pain. The fine level of granularity
means that problems can be detected when they
are very small and localized, and hence can be
dealt with before they get too large. Large-scale
pressure-sensing by massed arrays of micro-mechanisms
is being investigated for aircraft wings, where
detailed information about the pressure distribution
can provide useful information for the aircraft.
Learning
One of the key activities of living organisms
is that of learning. Activities like walking involve
fine control of actuators and large amounts of
sensory inputs such as vision and touch. Animals
are able to quickly recognize the meaning of signals
and the best way to control movements. The concept
of machine learning is the subject of current
research. Neural networks have recently been used
to learn touch patterns in a tactile sensing system
[Ref. 8].
Instinct
Living organisms often perform complex actions
based on instinctive knowledge – i.e., knowledge
that is somehow embedded in the DNA which is passed
on to offspring. The nest-making, song and migration
of birds are examples of actions that require
instinctive knowledge. Many birds make nests without
ever actually seeing how it is done, and many
birds fly on very complicated migratory routes
without ever having done it before. An instinct
such as migration involves the existence of detailed
information in the genes, such as star patterns
for navigation and instructions for the timing
of migration. The presence of instinct means that
creatures carry out their complex tasks without
any worry and with high levels of reliability.
There is no doubt that instinct is one of the
key reliability strategies used in nature.
In contrast to creatures with instinct, humans
must be trained in order to carry out complex
tasks such as operating machinery and constructing
equipment. The fact that humans are not constrained
to instinct can obviously have great advantages
in some situations. However, the fact that humans
need to be trained to carry out complex tasks
and that human behavior can be unpredictable means
that humans can often be one of the weakest links
in a complex system. Unpredictable behavior can
result from inadequate training, but it can also
be caused by personal problems or simply an adventurous
spirit. Nature does not suffer from these problems
because almost all of the activities of wild creatures
are controlled by instinct. To design a system
based on instinct ultimately means eliminating
the human operator. However, such a strategy would
appear to be unfeasible at present.
Simplicity in Nature
Nature provides powerful insight into how to achieve
simplicity in an optimum way in complex systems.
In nature, simplicity does not mean avoiding complicated
processes or even reducing the absolute number
of individual parts. In nature, simplicity is
maximized at the macro level by using a minimal
number of macro parts, a minimal number of sliding
parts, and continuous joints. These strategies
lead to a small number of simple macro interfaces
with the obvious benefits in reliability. To achieve
this extreme simplicity on a macro-level, there
is incredible complexity of functions on a micro-level
with very sophisticated smart materials. This
design philosophy in nature shows that simplicity
is a design goal that must be carefully defined
by design engineers.
Conclusion
The fact that so many reliability strategies can
be identified in nature demonstrates that strategies
can be an important means of achieving high reliability
in complex systems. Many of the reliability strategies
used by modern engineers are seen in nature. However,
there are several strategies used in nature that
are not used much in engineering systems. As engineering
systems take on more of the characteristics of
biological systems, there is potential benefit
in copying the reliability strategies of nature.
Acknowledgments
This work has been partly supported by an EPSRC
grant number GR/NO3648.
About the Author
Stuart Burgess is a Reader in Engineering Design
and is Head of the Design and Manufacture Research
Group in the Mechanical Engineering Department
at Bristol University, U.K. The Group is involved
in modeling the performance of mechanical systems
in engineering and nature. Research projects have
included spacecraft deployment systems and heavy
mechanical structures, along with studies in modeling
the performance of natural structures such as
trees.
References
- Roberts, M.B.V. Biology:
A functional approach. 4th Ed., Nelson,
p. 170, 1986.
-
Eliot, J.L. "Polar bears." In National
Geographic, Vol. 193, No. 1, p. 60, January,
1998.
- DEF STAN 00-41/3, "Reliability and maintainability
MOD guide to practices and procedures,"
p. 31-7, 1993.
-
King, A.S. and J. McLelland. Birds –
their structure and form. Bailliere Tindall,
p. 16, 1984.
-
Stoten, D.P. and E. Gomez. "Adaptive Control
of Shaking Tables Using Retrofit Strategies."
Proc. IASTED Conference on Control and Applications,
Banff, Canada. 1999. p. 402-407.
-
White, S.R. "Autonomic healing of polymer
composites." In Nature 409, p. 794-797,
2001.
- Dry, C. "Procedures developed for self-repair
of polymeric matrix composites materials,"
Composites Structures 35, p. 263-269,
1996.
-
Stone, R. "A distributive tactile sensing
technique for soft deformable contact."
Ph.D. thesis, Bristol University, 1997, xvi.
0569221.
|
