Page 34 - Winter 2020
P. 34

beam that are well suited for encoding the direction of a target by virtue of simple dependence between target direction and the signal frequency with the most signal energy (Müller et al., 2008).
Evolution Setting Ears in Motion
As if the shapes of the noseleaves and pinnae of bats were not complicated enough already, many bat species also have the ability to move these structures under muscular actuation. This insight from the biology of bats has triggered a next stage in the evolution of bat robots where the robots started to not only look like bats but also to have some of the animals’ noseleaf and pinna mobility.
The first bat robots with this kind of dynamics were able to carry out rigid pinnae rotations where the robots could change the orientation of their pinnae (Figure 5). The geometries and hence also the acoustic properties of the pinnae did not change during the reorientation; they were just orientated in a different way in the surrounding space. Such rigid rotations of the pinna can be used to determine the direction of a target. For example, by monitoring the received echo amplitudes across a rotational sweep, it is possible to determine which direction a target is in. One just needs to take note of the pinna orientation at which the maximum echo amplitude occurs, which should coincide with the direction of the target (Walker et al., 1998).
Rigid motions of a biomimetic sonar can consist of rotations as well as translations (Figure 6A), and their use is not limited to determining the direction of a target. Another important application is identification of a target. If the echoes returned from different targets differ only in very subtle ways, a special effort is necessary to bring these small distinguishing features to the fore. This can be expected to be true in the sensory worlds of bats that hunt their prey amid dense vegetation.
An example from the sensory world of bat robots can be found in the classification of the head and tail sides of a coin (Kuc, 1997). The relief of a coin is just a fraction of a millimeter thick, whereas the ultrasonic wavelengths typically used by bat robots are several millimeters long. Hence telling heads and tails of a coin poses a daunting task to any bat robot. Mobility of a bat robot mounted on the end of a robot arm with five degrees of freedom that included ear rotations (Figure 6B) has been used to enhance the signal- to-noise ratio by positioning the target coin in the center of the beam. Furthermore, positional errors could be taken into account by virtue of a set of echo templates that were formed by scanning a target in elevation.
The Rise of the Soft Robots
The things that real bats can do with their noseleaves and pinnae go far beyond the rigid rotations that have been mimicked by the bat robots of the 1990s and early 2000s. Bat species with sophisticated biosonar systems such as the horseshoe bats (family Rhinolophidae) and the closely related Old World roundleaf bats (family Hipposideridae) also have elaborate musculatures on their noseleaves and pinna.
For example, a single pinna in a horseshoe bat contains slightly more than 20 muscles (Schneider and Möhres, 1960). Furthermore, many of these muscles are arranged
in a way such that muscular contraction causes a change in the pinna shape. It should also be noted that these shape changes happen very fast. For example, the pinna of a greater horseshoe bat (Rhinolophus ferrumequinum) has been found to transition from one extreme shape configuration to the other in about one-tenth of a second.
A human blink of an eye takes three times as long.
What has driven the evolution of bat pinnae to such extremes in intricacy and speed? Because the pinna is in charge of diffracting incoming sound toward the entrance
 Figure 5. Bat robot with two rotational degrees of freedom for each receiver (ears; top). Rotation torques are transmitted to each receiver via a belt-and-pulley system to keep the rotating mass low. The stepper motors driving the ear motions are mounted below.
 34 Acoustics Today • Winter 2020

   32   33   34   35   36