Chapter 3: Figure and Ground
50 - Fly Spies
Scientists elsewhere are looking at how flies manage to fly. At first glance they shouldn’t be able to. The aerodynamic lift forces used by planes and birds don’t work for small insects. They have relatively large bodies as compared with their wingspans. And at their scale, the air’s viscosity makes the act of flying more like swimming in molasses.
With that in mind, Michael Dickinson of the California Institute of Technology suspended a bird-size pair of model insect wings covered with sensors in a drum of mineral oil. What he and others doing similar work discovered has interesting parallels with the MIT fish tail.
When an insect sweeps its wings forward and back, at the end of its return stroke it rotates each wing to create a backspin. This causes lower pressure above the wing and a momentary lift. Then, as the wing sweeps forward, it creates a rolling “leading-edge vortex” on top of the wing, which provides more lift. And at either end of the stroke, the wings also play off those front and back vortices in much the way a fish’s tail presses against the whirlpools it creates. It’s by using these tricks that insects can do things like take off backward and land upside down.
More remarkable, flies do all that with brains the size of poppy seeds. “A fly has, on average, 350,000 neurons,” Dickinson notes (we have 100 billion). “Most of them [are] dedicated to processing sensory information.” As much as three-quarters of its brain is devoted to just the eyes. Yet “when a fly breaks out of its pupa, it can fly as well as it ever will.” Dickinson is interested in how the fly’s brain works, he says, in how it’s possible for relatively few neurons to produce such complex behavior. “Brains evolved integrally related to bodies,” he notes, “and bodies evolved in the physical world.”
The Caltech team relies heavily on computers for modeling and tracking its flies. They also use an exotic array of computer-linked machines with names like Robofly (the fly simulator), Fly-O-Rama (flight path recording chamber with video inputs), Rock-n-Roll Fly Arena (flight recording chamber that pitches and yaws, with variable light arrays), FlyBall (neo fly eye), and the perhaps inevitable Bride of Robofly.
Dickinson is working with Ronald Fearing—a colleague from the University of California at Berkeley—on the development of a micromechanical flying insect. They have fabricated a diminutive robot thorax, and wings that will beat 150 times a second. A prototype that can hover in still air is their goal, but that’s only the start. The possibility of roboflies was first raised in a 1992 Rand Corporation study for the Pentagon. Several million dollars in funding for their development has now been awarded to the Dickinson-Fearing group and to such places as the Georgia Tech Research Institute in Atlanta. Robofly funding comes from the Office of Naval Research and the ever-present DARPA—the Defense Advanced Research Projects Agency.
The military envisions a new class of reconnaissance device called a MicroFlier, which individual personnel can carry and release. Once the operator gives a robofly general instructions about where to go, it will quickly disappear from sight and complete the mission by itself. It will be programmed to compensate for obstacles and wind drift on its own.
Email this to a friend | 
Add to del.icio.us | 
Add to digg |
Linkology by Kate (all links are unofficial) | Rights Info
