San Francisco Chronicle, June 24, 2002

LINK

Soaring achievement in spying UC Berkeley team creating 'microfly' to infiltrate enemies

Chuck Squatriglia, Chronicle Staff Writer

Looking to nature to improve on engineering, UC Berkeley researchers have cleared a major design hurdle in their quest to build a tiny robotic fly that the Defense Department hopes to use as a spy.

The robot, called "microfly," remains a long way from buzzing around a room,

but the recent creation of minuscule wings that flap like their biological counterparts marks a breakthrough in understanding how insects fly, scientists said.

Microfly is at the vanguard of a field known as biomimetics, based on the idea that biological systems are more flexible and adaptable than anything coming out of a laboratory. Efforts to reverse-engineer nature are influencing fields ranging from jet propulsion to submarine design.

The Pentagon is among the biggest supporters of biomimetics. The government has poured millions of dollars into microfly and other projects -- including a robotic lobster -- that it believes could yield new tools for warfare and espionage.

Microfly "can be used for reconnaissance and scouting," said Promode Bandyopadhyay, head of robotics at the federal Office of Naval Research. "You could have a swarm of them in a battlefield. Eventually, they can work as a group and detect the presence of hostile forces and materials."

Microfly will stand about a half-inch tall and weigh 100 milligrams, or slightly less than a paper clip. It will zip along at 3 meters (about 10 feet) per second and have a range of about 2 kilometers (about 1 1/4 miles).

Such a diminutive robot would be perfect for exploring caves, spying on terrorists or locating survivors in the rubble of an attack or natural disaster, Bandyopadhyay said.

But why a fly? Because the aerodynamic principles that allow F-14 jets to slice through the sky do not work on so small a scale. Instead, engineers must look to insects, which are robust flyers capable of aerobatic maneuvers that would tear a fighter jet apart.

RESEARCHERS MIMIC FLIES' WING MOtions

Microfly was hatched in early 1998, when the Office of Naval Research sought ideas for small robots. Ron Fearing, a UC Berkeley electrical engineer who has been fascinated by robots since childhood, suggested microfly. The government loved it and allocated $2.5 million.

Trouble was, Fearing had no idea how to make microfly fly. Cal biologist Michael Dickinson helped solve that riddle in 1999 with the discovery that insects use a complex choreography of three different wing motions to generate lift and thrust.

That question answered, Fearing and a team of a dozen researchers and students set about designing wings capable of mimicking those movements.

The result was wings 10 millimeters long, 3 millimeters wide and 5 microns thick -- about one-twentieth the thickness of a sheet of paper. They are made of a cellophane-like material called polyimide and are so light that a single drop of water can render them too heavy to move.

GETTING ROBOT'S WINGS TO FLAP, Rotate

Designing the wings was easy compared with making them work properly. The trick was getting them to flap and rotate simultaneously at 150 beats per second -- the precise speed needed to generate lift and thrust.

"We could get them to flap 150 times a second or rotate 150 times a second, but not both," Fearing said. It was frustrating, because "until you get the wings moving the right way, you can't do anything else."

The answer was a mechanism that allows the wings to flap and rotate simultaneously. A hinge-like linkage connects each wing to a piezoelectric motor, which is simply a ceramic strip that flexes under an electrical charge.

Microfly will rely upon the sun for power, and its three solar panels will double as landing gear. It will be fitted with electronic gear such as a tiny camera and transmitter, and a microprocessor with an operating system dubbed TinyOS will control it all.

Everything will be mounted on a stainless steel thorax, or body. Because microfly is so small -- it can sit atop a quarter with room to spare -- it cannot be assembled with nuts and bolts or heavy glues. Instead, everything is folded and linked like origami.

"It's not tedious so much as time-consuming," said Srinath Avadhanula, a mechanical engineering doctoral student who has spent hours assembling microfly parts under a microscope.

MAKING MICROFLY NAVIGATE PRESENTS Hurdle

The wings don't yet generate enough lift to get microfly airborne. Fearing hopes to solve that problem soon and begin work on a guidance system that will allow microfly to fly independently. It now flies only when tethered to a thin rod.

"It's a fly with training wheels," he joked.

Fearing is counting on Dickinson to help design a navigation system. Dickinson is studying how flies navigate with reaction speeds that allow them to change course in just 30-thousandths of a second.

Dickinson subjects flies to a battery of flight simulators that let him manipulate the scenery and measure their responses. He's discovered that a fly -- with a brain the size of a sesame seed -- can collect, process and respond to information with the efficiency of a supercomputer.

"The information is modulated before it even gets to the brain," he said. "It's quite complicated and elegant."

Microfly won't be the first product of biomimetics. Perhaps the best-known example is Velcro, which was modeled after the burrs that stick so tenaciously to clothing. But projects such as microfly represent a substantial leap in complexity.

"You have no idea how good biological systems are until you try to build a man-made one," Dickinson said. "Only then do you realize we have a lot to learn from how these things are put together."

NEXT UP: 'ROBOLOBSTER,' 'ROBOpike'

Among the projects receiving Pentagon funding is "robolobster," a $3 million project under way at Northeastern University in Boston. It will mimic a lobster's ability to navigate pounding surf. The Navy hopes to use robolobster to clear mines and other explosives from shallow water.

The Office of Naval Research also bankrolled "robopike," which is helping researchers at the Massachusetts Institute of Technology understand the fluid mechanics that let fish propel themselves almost effortlessly.

Biomimetic discoveries could bring tremendous advancements to the study of aero- and fluid dynamics, said Bandyopadhyay of the Office of Naval Research.

For example, understanding how flies produce the lift and thrust needed for flight could lead to stealthier jet engines. Learning how bats find insects in the dark could improve sonar and radar systems. And discovering how a pike can cover 6 meters per second with the flap of its tail could lead to faster submarines.

"The Navy is not trying to build a robotic zoo," Bandyopadhyay said. "Instead of building something that looks like a fish or a dolphin or a whale, we want to distill the science and apply it to existing platforms."

For more information about microfly, log on to http://robotics.eecs. berkeley.edu/ronf/mfi.html.

E-mail Chuck Squatriglia at csquatriglia@sfchronicle.com.

-- (lars@indy.net), June 24, 2002

Answers

not the Chuck Squatriglia?

-- (lars@indy.net), June 24, 2002.