Step inside the Laboratory of Intelligent Systems at the Ecole Polytechnique Fédérale de Lausanne in Lausanne, Switzerland. Here, in a country perhaps best known for its fine chocolates and wristwatches, engineers are building awesome robots with formidable powers. But you may have to squint a little to see them at work, because the Swiss bots can be kind of hard to see.
“Each is about the size of a locust,” explains Miko Kovac, the machines’ creator.
Unlike the thundering, lumbering monsters that staggered through several decades’ worth of science fiction novels, comic books, and movies, these 21st century robots are so small, they can slip between a user’s fingers—or pass through the eye of a needle. Like circuits and SUVs, robots are getting downsized. The result is a rapidly growing number of handy helpers designed to accomplish tasks in medicine, security, and several other fields that are too difficult or dangerous for people or even larger robots to tackle.
“We’re making robots that are tiny, nimble, and fearless,” Kovac says.
Yet as robots get increasingly minuscule and versatile, their creators are encountering challenges in controlling, moving, and powering their creations.
“When you get down to the microbot level, many conventional technologies go out the window,” says Randall Shumaker, director of the Institute for Simulation and Training at the University of Central Florida. “In many different areas, because of the size scale involved, you have to come up with entirely new ways of doing things.”
Microbots in Motion
Like other microbot developers, Kovac believes the technology’s greatest potential lies in its ability to explore places that are too cramped or otherwise inaccessible to humans, animals, or conventionally sized robots. But simply getting tiny bots into the places where they are needed is sizing up as a major challenge for developers, since intricate walking- and wheel-based locomotion technologies are difficult to achieve on a micro or smaller scale.
Microbot researchers are addressing this problem with a variety of different and innovative approaches. Kovac’s solution was to develop a bot that jumps like a grasshopper (Fig. 1). His prototype weighs just 7 g, but it can leap 1.4 m—more than 27 times its body size.
“That’s 10 times farther for its size and weight than any existing jumping robot,” says Kovac, who is developing the robotic bug as his PhD project. He observes that his “biomimetic” jumping technology is unique, since it allows the microbot to travel over the kinds of rough terrain that would block most walking or wheeled devices.
Small jumping animals, such as fleas, locusts, grasshoppers, and frogs, use elastic storage mechanisms to slowly charge and quickly release their jumping energy, Kovac explains. In this way, they can achieve very powerful jumps and very high accelerations. Kovac’s jumping robot uses the same basic principle, charging a pair of torsion springs via a small 0.6-g pager motor and a cam.
To be able to optimize the jumping performance, the legs can be adjusted for jumping force, takeoff angle, and force profile during the acceleration phase. Kovac notes that his bot could be fitted with tiny sensors to allow it to explore rough, inaccessible terrain or to aid in search and rescue operations.
Duke University researchers are also taking a mechanical approach to microlocomotion, though on an even smaller scale than Kovac’s machine. The Duke engineers are developing tiny bots that are designed to roam inside a laboratory-on-a-chip to perform various experimental and housekeeping tasks. Each microrobot resembles a kitchen spatula, but with dimensions measuring just microns (Fig. 2).
“We’re talking about robots that are invisible to the unaided human eye,” says Bruce Donald, a Duke computer science and biochemistry professor.
Helping the microbots get around, a “scratch-drive” motion actuator enables the devices to propel themselves across surfaces in inchworm-like fashion. The microrobots advance in steps measuring only 10 to 20 billionths of a meter. On the other hand, their pace is blazingly fast.
“Little steps can add up very quickly, especially when they’re repeated at up to 20,000 times per second,” Donald says.
Cornell University researchers are also looking to nature to uncover ways of moving microbots quickly, accurately, and efficiently. Working on a scale that makes using conventional electronics impossible, the team has turned to biological chemistry for inspiration and is looking at sperm cells, of all things, as a potential propulsion model.
Sperm cells move with the help of a wiggling flagella, a tail-like structure that propels the cell forward (Fig. 3). Alexander Travis, a Cornell assistant professor of reproductive biology, believes that the same chemical energy that moves sperm cells can be used to drive nanoscale robots.
“We’re interested in creating an energy delivery to power nanoscale robots,” he says. “The approach works well in nature, so there’s no reason to believe that it won’t work with nanoscale robots.”
Travis sees potential applications in medicine, including delivery systems loaded with chemotherapy drugs or antibiotics to target specific cells. Such a system would allow doctors to provide steady doses while reducing side effects that result from treating the entire body with a drug.
Getting microbots to move is crucial, but so is controlling their actions. In fact, perhaps the biggest challenge currently facing microbot developers is finding effective ways of controlling their creations, Shumaker says.
Radio control, the preferred method of directing and coordinating robot operations, becomes problematic when machines are scaled down to toy-sized or even smaller dimensions. Depending on the system’s size and shape, it can become difficult—sometimes even impossible—to add basic wireless technology to a microbot.
“Even if you do find a way of squeezing a receiver onboard, and the microbot can handle the added weight, you face the problem of installing an antenna that’s large enough to provide decent reception capabilities,” Shumaker says.
The Duke researchers are addressing control issues by designing their bots to respond to electrical stimuli as they move through a chip’s pathways. Built using microchip fabrication techniques, the bots are each designed to respond differently to the same single “global control signal” as voltages charge and discharge on their working parts.
The idea has paid off in rather spectacular fashion. Donald’s team recently got five of the microbots to group-maneuver in cooperation under the same control system. During a recent demonstration, two of the devices were directed to pirouette to a Strauss waltz on a dance floor measuring just 1 millimeter across. In another exhibition, the bots were directed to pivot in a precise fashion whenever their boom-like steering arms were drawn down to the surface by an electric charge.
“The response is similar to the way dirt bikers turn by extending a boot heel,” Donald says.
Although microbot researchers are developing a variety of impressive control methods, Shumaker believes that the ultimate and definitive answer to microbot management lies in autonomous operation.
“We have to at least downplay the notion that people will be able to continuously direct microbot operations,” he says. “Autonomous operation allows a microbot to decide for itself where it will go and what it will do, basically eliminating the need for continuous human oversight.”
Autonomy becomes a necessity when deploying swarms of flying or rolling microbots.
“It’s one thing to manually direct a very large robotic drone, like the Air Force’s Predator,” says Daniel Burrus, CEO of Burrus Research, a technology trends forecasting firm. “But it would be very difficult for a human, or even a team of people, to coordinate the interaction of hundreds or even thousands of tiny moving robots.”
Fortunately, a new generation of smaller and more powerful processor and memory devices is enabling microbot developers to add a great deal of intelligence to very small systems. The biggest remaining roadblock lies in developing the sophisticated onboard software required to manage microbot activities.
“Software is the key element and is becoming one of the most active areas of research,” Shumaker says.
Autonomy is already beginning to appear in some microbots. Researchers at the Israel Institute of Technology, for instance, have developed a neurosurgical microbot that directs itself using detailed information obtained from electronic preoperative patient scans.
During surgery, the microrobot is affixed to a head clamp attached to the patient’s skull. The bot automatically positions itself with fine accuracy, locks itself in place, and then serves as a guide for the insertion of a needle, probe, or catheter by a surgeon to carry out a specific procedure such as removing a tumor.
The technology helps reduce the amount of pain a patient experiences, since it enables faster surgery by allowing the surgeon to know exactly where to insert the surgical instrument, says Moshe Shoham, director of the school’s robotics laboratory and one of the microbot’s developers. Surgeons generally like the system because it’s easy to use and is applicable to a wide variety of neurosurgical procedures. There are some other benefits as well.
“It allows much less invasive surgery \[and the need for\] much less radiation,” Shoham says.
While microbots have relatively modest power needs, most also have little spare room available for storing fuel or other energy sources. This has forced designers to think both small and creatively, turning to both proven and highly experimental technologies to satisfy their microbots’ power requirements.
Working on a limited budget, Kovac opted to use a tiny battery—a relatively mundane and low-cost technology—to power his jumping microbot. Since charging or replacing a microbot’s battery would be difficult and time consuming, Kovac designed his system so it could be fitted with small solar cells.
“The cells help the robot to recharge between jumps,” he notes.
At Cornell, Travis and his team are targeting a couple of energy sources to power their sperm tail propulsion system. The researchers have zeroed in on glycolysis, a process that breaks down glucose to derive adenosine triphosphate (ATP), a substance cells use for energy.
The glycolysis process requires 10 enzymes. Using special “targeting domains,” sperm tether these enzymes to a fibrous sheath that runs the length of the tail. The researchers are trying to recreate this glycolytic pathway by modifying each protein’s targeting domain so they can instead bind to nickel ions onto a manufactured chip.
So far, the researchers have successfully attached three of the 10 enzymes required to make ATP from glucose. If they manage to attach all 10 enzymes, each enzyme will in principle act in a series to ultimately generate ATP to power a nanodevice. In the body, such a device could conceivably use readily available blood glucose as fuel.
In the years ahead, at least one family of microbots may be able to draw energy from the same materials that the bots are constructed from. This is the promise of chembots, the convergence of soft materials chemistry and robotics. Formed out of electro- and magneto-rheological materials and other exotic chemical combinations, self-powered chembots will literally ooze their way into otherwise inaccessible places.
Chembot researchers expect their systems will be able to perform feats unmatched by any current machine. The systems would be able to enter confined or complex spaces; follow cables, ropes, or wires; and climb trees or other branched structures.
Once in place, an energy self-contained chembot could survey an area and then morph into a specific form to accomplish its task. The device might, for example, gain entry to an improvised explosive device to gather information or potentially disable the device. Other applications include landmine detection, search and rescue in hazardous conditions, and biomedical diagnosis.
A New Era
As microbot research evolves and matures, developers are finding an increasingly deep well of design tools available for addressing critical issues.
“Depending on the situation at hand, a developer may already be able to select from two or three different approaches to a specific problem,” Shumaker says.
Additional advances are certain to come, Shumaker notes. In fact, future historians may someday look back at the early 21st century as a golden age of microbotic innovation and, perhaps, the dawning of an era as important as the introduction of the automobile, airplane, or the Internet. Burrus observes that tiny robots seem destined to play important roles in national defense, home and business security, medicine, transportation, home maintenance, and an almost endless number of other fields.
“As robots get smaller,” he says, “the number of possibilities seems to grow greater.”