Mechanical Mashups

Earth's early life forms come face-to-face with sophisticated technologies

Imagine a flying robot that could peek in on suspected terrorists; a tiny crawling machine that could root out survivors of a natural disaster or scout inside a burning building; or even a bitty bot that could creep through your artery and reinforce a weak spot.

Researchers at Case Western Reserve University and around the world are working daily to develop such machines. And these developments have something unusual in common: These innovations—the very future of robotics technology—are inspired by some of Earth's earliest creatures.

More likely to elicit eeks and ewws than oohs or aahs, bugs—the antiquities of the animal world—turn out to be surprisingly complex. The creatures easily navigate environments hazardous to humans and go into rough and cramped terrain where no person would, or in some cases could, ever step foot.

Cockroaches and the like creep and crawl; they burrow and skitter through tight spaces and rocky landscapes. Moths and other winged insects can fly with precision on a moonless night. Bugs are cheap. They're easy to raise, durable and—most important—they can execute the exact kinds of endeavors scientists, engineers and the United States military would like to hand over to capable robots.

Through hundreds of millions of years of natural selection, bugs have continually become better bugs, says Case Western Reserve University biologist Roy Ritzmann, PhD, who has been working on biologically inspired robots for 20 years.

"When you consider locomotion, what drives evolution is physics—the same physics robot-makers have to use," he says. "Why reinvent? Evolution has done the job for you."

Inspired Improvements

Roger Quinn, PhD, a mechanical and aerospace engineer, is head of the Biologically Inspired Robotics Laboratory at Case Western Reserve. He was among the first engineers in the country to tap into the powers inherent in creepies and crawlies. In collaboration with biologists like Ritzmann, he has built robots based on animals at the very bottom of the food chain for two decades. His lab floor and bench tops are littered with mechanical versions of cockroaches, crickets, moths and worms. "If you look at the possibilities, robots can save lives. But we need robots with a lot greater capabilities than we have currently," Quinn says. His work to create such sophisticated machinery has garnered interest among the military, NASA, the Department of Homeland Security, fire and police departments and medical companies.

Ritzmann points to the robots that were used for search-and-rescue efforts at the World Trade Center after the Sept. 11 attacks. The effort, he says, revealed a major shortcoming: they were unable to navigate autonomously—or even semi- autonomously. Operators above ground had to steer robots over or around every obstacle. Spotty communications caused further troubles.

"We want to develop something akin to a well-trained dog," Ritzmann explains. "We want to be able to throw the ball and say, 'Fetch' but not have to say, 'Go over that rock and turn left, but avoid that shelf.' "

That doesn't mean, however, that he and others can mirror exactly what nature has taken centuries to devise. Instead, Quinn says, the idea is to be inspired by the abilities insects have developed.

"Typically, you're better off not trying to mimic animals because you don't have the same materials to work with," he says. "Instead of parroting them, you learn from the animals, and you build their capabilities into the robots."

Deceivingly Complex

As early as 1990, Case Western Reserve biologist Hillel Chiel, PhD, and then-Case Western Reserve computer scientist Randall Beer, PhD, suggested that truly flexible and adaptive abilities would emerge from robots whose bodies and brains were inspired by real animals, even relatively simple ones. The approach has caught on across the globe.

"In the last 10 years, robots have come out of the lab and shown they can function outside," says Mark Cutkosky, PhD, a mechanical engineer at Stanford University who has built some of the world's fastest-running robots based on insects.

In Japan, a group of researchers recently produced a flying artificial butterfly. Engineers from South Korea built a robot that swims like a jellyfish. The locust is a model for jumping and flying robots in Israel and Switzerland; spiders and crabs have influenced walking bots in Spain.

Back in the states, University of Michigan's Khalil Najafi, PhD, fit a beetle with devices collecting power from the sun, movement and body heat to help power sensors. At Harvard University, Robert Wood, PhD, mass-produces mechanical flies using origami and pop-up book technology. He wants swarms to pollinate crops if bee populations continue to crash.

But to make the robots practical, Cutkosky says, still better power and a substitute for muscle, which provides movement, structure and energy storage, are needed— a reminder that the world's oldest creatures are more complicated than most people give them credit for.

"You have to be careful calling any animals simple, because they never are," says neurobiologist Barry Trimmer, PhD, of Tufts University, a leading builder of soft- bodied robots. His are based on caterpillars.

In a roach, fly or worm nervous system, neurons multitask, which makes the creatures pretty complex. Still, these heavy-hitting cells come in comparatively manageable numbers, Trimmer says. "The human brain has hundreds of millions of neurons; a fly brain has 200,000. A caterpillar has 70 motor neurons—one nerve cell controlling each muscle—compared to 50 million neurons in each arm of an octopus. In a fly or a caterpillar, you can more easily, in theory, find what the brain is telling the body."

Learning this connection between brain and body was the goal for Ritzmann's research assistant, Alan Pollack, who earned his MBA from Case Western Reserve in 1990. Pollack spent a year perfecting surgical techniques for probing the roach brain. He then delicately cut through the bug's brain sheath and inserted a hair- thin braid of four wires to monitor activity. His experiment showed the tiny area of the cockroach brain that processes information from the outside world also fires neurons down to local control-centers at each leg, signaling when to walk, run and turn. Ritzmann's and Quinn's labs now are testing an artificial nervous system connected to a mechanical cockroach leg.

Inching Forward

Building robots based on seemingly simple creatures, however, is anything but easy. Chiel and Quinn point to their long-standing effort to build an earthworm robot. They first described their soft, underwater worm-like machine in a 2000 study. Doctoral candidate Alexander Boxerbaum, who studied art history as an undergraduate student, recently helped them see the animal's movement in a new light.

"Engineers try to break down what they see in nature into the simplest elements," Boxerbaum says. "An earthworm has more than 100 segments, and everyone who has tried to make a robotic worm has reduced that to five or six. We had oversimplified."

A worm's movement relies on a smooth wave, Boxerbaum says, and their new robot—made of smooth mesh—crawls faster and doesn't slip like earlier models. Chiel acknowledges the robot could one day act as a spy that creeps through pipes of a suspected terrorist's apartment, but it likely holds even greater potential in performing day-to-day operations.

"A worm-inspired robot could be made to fit inside and inspect water mains, finding damage before a road collapses," he says. "Smaller versions could be used for colonoscopies or scaled down further and used as a stent that would crawl though blood vessels."

To be a Moth on the Wall

Spying, though, remains top-of-mind for biology professor Mark Willis, PhD, and his lab mates. The military, which funds their investigation of tobacco hornworm moths, is interested in creating fly-on-the-wall-sized robots.

While fixed-wing planes become unstable at such a small scale, "These moths, which have a 6-inch wingspan, are some of the most capable flying machines Mother Nature has created," Willis says. "They can fly fast, climb, turn and dive, hover and fly backward. They're like insect helicopters."

Willis works with U.S. Air Force research labs in Ohio and Florida, which have called for researchers to take a good look at the basic science behind flapping- winged flight. Ultimately, the idea is to build robots that can fly through a window and identify who's inside, or go see what's around a corner, saving soldiers from potential threats.

With researchers at the University of Washington, Willis found the antennae, which male moths use to track odors of females, also act as a sort of gyroscope. Cut half an antenna off, and the moth bobs and weaves like a drunken driver. Glue the antenna back on, and the moth flies normally.

"These insects have multipurpose sensors, just what you want when space is tight," he says.

Taking the cue from the moth's antennae, Adam Rutkowski, PhD, a doctoral candidate at the time, developed an artificial algorithm that allows a flying robot to estimate its own motion, wind velocity, ground speed and flight height using just two sensors.

Powered Up

Rather than develop an entire robot from scratch, Case Western Reserve chemist Daniel Scherson, PhD, is trying a different route with his studies—incorporating robotics in bugs themselves.

He has controlled live insects by triggering their nerves with electrical impulses. The plan, he says, would be to outfit an insect with a tiny camera, recording device or sensors. It would maintain communication with its handlers to relay back what it encounters on its journey. This would, of course, require a power source. And, like full-fledged robots, Scherson says, such a cyborg would need an alternative to today's batteries.

So he, along with then-doctorate student Michelle Rasmussen, PhD, and Case Western Reserve chemist Irene Lee, PhD, devised a tiny fuel cell that turns sugar in a cockroach's blood into electricity.

The cockroach can generate its own power as long as it eats normally.

In a recent test, the cell powered a nickel-sized, low-frequency radio transmitter designed by Case Western Reserve electrical engineer Steve Garverick, PhD, and master's student Bill Weeman. A few inches away, a radio receiver picked up the signal. Scherson's lab has miniaturized the biofuel cell to fit fully inside a cockroach and, with Ritzmann and his senior research assistant Alan Pollack, attached the transmitter to the bug establishing wireless communication with the receiver. The same strategy was applied by PhD graduate student Jamie Schwefel for Willis' moths with the same results.

"We've got a long way to go," Scherson says. "But these are important steps."