Robotic Bug Project Takes Flight After Looking to Nature for Design Ideas

It sounds like a scene in a really bad movie: Thousands of tiny mechanical bugs crawling across the landscape, looking for evil.

But that’s exactly what researchers at Vanderbilt University hope to produce with two grants from the Defense Advanced Research Projects Agency.

Ephrahim Garcia and Michael Goldfarb, mechanical engineering professors at the Nashville university, are designing tiny robots about a third the size of a credit card that could be dropped from airplanes or released by soldiers in the field.

The bugs, as Goldfarb calls them, would literally “swarm” across the land, traveling up to five-eighths of a mile and looking for everything from mines to chemical weapons to enemy soldiers.

Under a $904,000 three-year grant, the researchers hope to produce their first crawler by next summer. But the project recently got a lot harder with a second DARPA grant for $780,000. The agency isn’t satisfied with crawlers: It also wants the insects to fly.

That means the motor-actuated robots of the recent past won’t do. The new insects must borrow a few techniques from nature, and they have to work with no more power than the battery that runs an ordinary wristwatch.

So the researchers abandoned the concept of motors that drive moving parts in favor of the system used by real insects: vibration and deformation.

In the preliminary design of the bugs, “everything moves by deformation, so things just bend and flex a lot,” Goldfarb said. A small electric current will be applied to thin ceramic-coated metal plates that bend when the current flows and snap back to their original form when the power is off. Similar “piezoelectric” actuators make a pager vibrate.

The researchers are designing four- and six-legged versions. The actuators will shake the entire structure, causing it to move in such a way that the legs “vibrate in a walking pattern,” Goldfarb said.

The insects will be largely autonomous, but subject to simple radio commands to move forward or turn right or left. Without electric motors, the bugs will be far more energy-efficient, according to Goldfarb’s colleague, Garcia.

“An electric motor is continually running,” Garcia said. “We’re pulsing energy in as we need it.”

But even if the researchers are successful in producing crawling electromechanical insects, how are they going to make them fly? Goldfarb said they turned to the biological community for help.

“Vibration is the way real flying insects work,” he said. “Years ago, people looking at insects found that the wings of an insect flapped five times faster than its brain could send the impulse. They didn’t understand how that could be.”

It turns out, he said, that the wings are just an oscillator.

“All the brain was doing was sending an impulse to that oscillator every fifth motion,” he said.

He likens it to pushing a child on a swing. It isn’t necessary to push the swing each time. Once every five times will keep the swing moving.

Similarly, a tiny electric pulse could keep the insect’s oscillator in motion so its aerodynamic wings continue flapping.

That, said Goldfarb, would allow the insects to both crawl across terrain and fly over tall grass or even a canyon to continue their reconnaissance.

The Vanderbilt team is not designing the sensors to ride aboard the bugs, but Goldfarb said a number of other existing devices are small enough and energy-efficient enough to do so.

“A magnetometer, something that could detect metal, would be very useful,” he said. “Usually, metal is bad. When your soldiers find there is metal sitting over the next hill, it’s usually not a good thing.”

Sensors that detect the presence of harmful chemicals or mines are another option.

The bugs will be so small that, most likely, each will carry only one type of sensor. Goldfarb sees them being deployed in swarms of hundreds, if not thousands.

“The swarm could spread out and look for things,” he said. “Each would know to some extent where its nearest neighbors are. So if one finds something interesting, and it needs some sort of verification, it can call other bugs with different sensors over to investigate.”

In the future, some of the bugs might carry cameras, but Goldfarb said that is not yet in the works. It will be difficult to design a camera within the size, weight and energy parameters of the system, he said.

The goal is to produce the bugs so cheaply that they will be expendable. That way it won’t be necessary for soldiers to return to the field to get them, he said.

Goldfarb is the first to admit there may be formidable obstacles down the road.

“It’s a very new topic, and there aren’t a lot of precedents, so it’s hard to say exactly what we will be able to do,” he said.

It’s an interesting and challenging concept, but it may turn out to be harder to get the bugs out of the design than to get the insects into the field.