Engineers Look To Nature For New Robotics Strategies

By Alex Compton

If you're like most people, the word "robot" invokes the image of a heavy-set, metallic humanoid with telescoping arms and an array of non-descript buttons and flashing lights across its chest. Our notion of robotics is undoubtedly influenced by early science fiction writers of the mid-20th century, who envisioned futuristic autonomous machines as improvements made upon the human form. This may be a reflection of our selfish tendency to view human beings as evolutionary pinnacles: highly optimized creatures that represent the most advanced life form on Earth.

It's not surprising then that the human body was initially used as a model for the advanced robotics industry. In recent decades, however, engineers have shifted their attention elsewhere in an effort to make efficient robots that can be blasted into outer space or sent to the front lines of military conflicts. And as illustrated in the December 2009 issue of the journal Bioinspiration and Biomimetics, researchers at Oregon State University have found mechanical marvels in the most unexpected of places.

The common cockroach, Periplaneta americana, is one of the most ubiquitous creatures on the planet. While universally renowned as an unwanted pest in homes, hotels, and restaurants, the insect has gained the respect of biologists who study its ability to survive in harsh environments. Indeed, cockroaches are among the hardiest creatures on the planet. They have been shown to survive for months without food or water. Some can go without air for 45 minutes, even while submerged underwater. They are remarkably resilient to nuclear radiation, and as a result it is popularly suggested that cockroaches will "inherit the Earth" following a catastrophic nuclear war.

Now the notorious insect is the focus of study not by biologists, but by an unlikely group of scientists: academic engineers.

John Schmitt, an assistant professor in the School of Mechanical, Industrial, and Manufacturing Engineering at Oregon State University, is meticulously studying cockroach movement in order to design a legged robot that runs effortlessly over rough, unpredictable terrain.

Historically, engineering encompassed the design of inanimate objects that help humans overcome nature rather than emulate it. In recent decades, however, scientists and engineers seeking new technological solutions have looked no further than the awesome displays of innovation that are directly observable in nature: the plants, animals, and microbes that have adapted remarkable displays of biomechanical prowess after more than four billion years of evolutionary fine-tuning. Engineers such as Schmitt design new technologies that mimic what's seen in nature, an approach known as biomimetics. This engineering strategy is being used by Schmitt to develop robots that act, and more importantly react, like insects.

"Inspiration can come from a variety of different animals; for example, cockroaches, guinea fowl and geckos all have remarkable locomotion abilities," says Schmitt. In the 1970s, Shigeo Hirose of the Tokyo Institute of Technology was one of the first to consider the slithering and the slimy as models for new robotic designs. His life's work has been dedicated to building robot prototypes that undulate like snakes.

Schmitt was smitten by cockroaches because they are particularly adept at scurrying over diverse landscapes while expending limited amounts of energy. He and his colleagues at OSU are interested in creating a running robot that not only moves with greater efficiency than humans, but also tackles rough, unpredictable terrain. The objective of the project is to identify mechanical principles inherent to the insect's biology that enable it to run quickly and effortlessly. They ultimately hope to use this knowledge to build robots that use minimal amounts of computing power while running across diverse landscapes.

"Cockroaches are incredible. They can run fast, turn on a dime, move easily over rough terrain, and react to perturbations faster than a nerve impulse can travel," Schmitt said in an OSU news release. In essence, cockroaches do not have to think about running, even when doing so involves adjusting to unexpected changes in ground surface. Cockroaches can automatically adjust to a terrain up to three times its hip height, without flinching. That's like a human stepping into a massive pothole without breaking stride.

"If a leg misses the ground during running, its muscle activation pattern remains the same; the leg continues to swing rearward and muscles are activated in the same manner as if the leg were touching the ground," says Schmitt. These observations suggest that cockroaches do not use reflex control to modify their running patterns. It seems that the structural framework of the insect endows it with such an uncanny ability to locomote. "Evidently, how the roaches are built is capable of conferring stability," he notes.

Schmitt hopes to translate his designs to a full-size running machine that can perform difficult jobs that aren't well suited for humans. "Robots that do not have to utilize so much processing capacity just to run could instead utilize [it] to perform path planning or coordinate with other robots. Robots that could perform in this fashion would have any number of uses," he says. His designs may produce mechanical rovers that move naturally over lunar terrain, search and rescue robots that could help locate victims of natural disasters, or robots that could be used for military reconnaissance missions.

Schmitt is trained as a mathematician and heads the theoretical side of the studies, while his collaborator, Johnathan Clark of Florida State University, deals with the biomechanics and robot design. Schmitt employs computer models to create robot prototypes that can automatically adjust to changes in terrain almost as well as the cockroach. Simulations allow his research team to determine how mechanical factors such as energy storage and leg angles are associated with increased running stability.

While cockroaches can tackle nearly any type of terrain at high speeds, they earn few style points for how they do it. Professor Robert Full and his research team at the University of California at Berkeley examine how cockroaches switch from running across flat surfaces to climbing a wall. In January, at the annual meeting for the Society for Integrative and Comparative Biology held in Seattle, his group revealed that these insects' locomotion is anything but perfect. High-speed cameras captured how the roach rams into the wall like a torpedo before making the vertical climb. "Instead of using their head and associated sensors for neural control, they used their head like a car bumper," he reported at the meeting. "They run into the wall headfirst and run up [it] like nothing happened."

But just because the insect is clumsy doesn't disqualify it as an engineering marvel. Ultimately, both research groups converge to emphasize that structural robustness is critical to cockroach locomotion. It turns out that the insect's strength and versatility make up for its awkwardness.

Schmitt and Full are not alone is their quest to find "bioinspiration" from common critters. Kristi Morgansen of the University of Washington is tapping the biomechanical potential of fish to design swimming robots that can be used to for underwater exploration and reconnaissance. Just as the cockroach is well adapted to run across sketchy surfaces, there are certain structural characteristics inherent to the fish that help it propel through the water efficiently. Morgansen, an associate professor in the department of aeronautics and astronautics, is hoping to imitate these fishy traits to make a more graceful underwater robot. "Devices with flapping fins as opposed to rotating machinery are less likely to get fouled in things like kelp," she says.

While the rationale for her fish-centered approach goes without saying, she studies more than just how fish like the tuna get from one place to another.

"In particular, communication underwater is particularly challenging. We're starting to look at how fish handle communication to see what we can learn," Morgansen said. Understanding how fish cooperate with one another as they swim may enable her to design robots that work "in sync" with one another.

Has the human form become obsolete in the field of robotics? Not necessarily. There are still plenty of roles to be filled by mechanical versions of ourselves. For example, Yoky Matsuoka from the University of Washington is designing robotic replicas of the human hand for use as a novel prosthetic. Blake Hannaford, also from UW, is working on robots that can automate surgery. In these cases, it seems like the human form is best suited for the job.

While biomimetics hasn't turned the robotics industry on its head yet, there's no doubt that the solutions to many engineering problems may be hidden beneath a tuft of fur, a thick exoskeleton, or a scaly skin. As Danish physiologist August Krogh once said, "For such a large number of problems, there will be some animal of choice on which it can be most conveniently studied."

Alex Compton is a graduate student in molecular and cellular biology at the University of Washington.

Image: Researchers are studying how cockroaches move effortlessly over unpredictable terrain. Photo: Robert Full