Biomimetic Self-Organization and Self-Healing - Engineered Biomimicry - page 330

information is transmitted via individual interac-
tions. In most biological and physical systems,
the correlation distance is significantly longer
than the interaction distance but shorter than the
size of the system. However, in starling flocks the
correlation distance is as large as the group, inde-
pendent of the number of individuals in the
group; the behavior of all birds is correlated,
which is known as scale-free correlation [8] .
Computer-automated measurements of indi-
vidual starlings in flocks ranging in size from
500 to 2,500 birds have revealed that the speed
of the entire flock (center of mass) averages
between 9 and 15 m/s. Perhaps somewhat sur-
prisingly, the flocks are not as dense as they
appear to a human observer standing on the
ground. The density is 0.04-0.8 birds per cubic
meter (compared to molecular density of materi-
als, this would be closer to a gas than a liquid)
and the distance from a bird to its nearest neigh-
bor ranges from 0.7 to 1.5 m [7] .
The nearest-neighbor distance in fish schools
is not independent of the size of the group, as it
appears to be in bird flocks [5] . Apart from this,
the collective behavior of fish schools closely
follows the behavior outlined for bird flocks.
The individual fish interacts with neighbors by
showing behavioral matching (i.e., it matches its
behavior with that of its neighbors) and posi-
tional preference (i.e., it maintains a constant
distance or position relative to its neighbors) [1] .
Similarly to bird flocks, this is thought to confer
fitness advantages to individual fish through
protection from predation.
This insight has potential for many applica-
tions. Computer animations use swarm algo-
rithms for simulating flock behavior among
birds, bats, herds, and fish at sea [4] . An interest-
ing observation is that human beings also exhibit
similar flock behavior when we crowd together—
for instance, when driving cars. The distributed
algorithms are therefore also usable for predict-
ing and controlling car traffic [4] .
Similarly, the field of robotics benefits from
studies of insects and other animals. Especially,

insects attract interest since they perform diffi-
cult tasks even though they only have very small
brains. This is important for designers of mobile
robots, because processing capacity is restricted
by weight and limited energy access. The interest
also goes in the opposite direction, where robots
can be used by biologists and engineers in col-
laboration to simulate the behavior of animals in
order to test a hypothesis explaining that behav-
ior [9, 10] . Robot researchers study and mimic
various capabilities in insects by looking at
chemical, visual, and auditory sensing and on
how complex motor control is performed [11] .
Robots are made that navigate like bees and ants,
which use visual landmarks to remember the
trail back to the nest. One example is the mobile
robot Sahabot 2 that uses a 360° camera that cap-
tures how the light intensity in a 360° view
changes during movement [11] . Other examples
are provided in the accompanying chapters on
microflyers (Chapter 5) and biomimetic vision
sensors (Chapter 1).
Biological inspiration from flock behavior is
also valued in bigger robots such as automated
cars. The decentralized control of unmanned
vehicles to form convoys, e.g., on highways, is
called platooning . Inspiration comes from birds
such as geese that fly in formation. The advan-
tage of platooning is that car drivers can relax
and do activities other than driving. Further-
more, it makes traffic safer and allows for higher
capacity on the roads since the distance between
the cars can be reduced and the risk of queues
due to variation in speed can be lowered. The
European Sartre project has recently demon-
strated platooning on a public road with eight
cars where only the front car had a human
driver [12] .
Finally, the flock behavior in birds and fish
attracts attention from researchers developing
unmanned aerial vehicles (UAV), i.e., small air-
planes and helicopters [13-15] . As discussed in
Chapter 5 on microflyers, effort has so far been
invested in the development of hardware and soft-
ware platforms to make the vehicles fly and sense