March of the Roaches


Aside from their arguably deficient minds, insects are in some ways superior to the larger beasts. They can fit in tight crevices, they can survive radiation, and they can carry many times their weight. These advantages and the diversity of their locomotive methods have inspired a recent spate of robots. Moths, spiders, ants, centipedes, and scarabs have scurried and flown across many a lab table.

Some are even available in the toy store. But that’s the only place they’ll likely have a purpose. The robot bug has one major deficit: A battery big enough to power a more grown task would be prohibitively heavy.

Alper Bozkurt, a professor in the department of electrical and computer engineering at North Carolina State University, has a solution. If you want a bug-like thing to work for you, forget the robots. Use a real bug. “Our aim was to solve this problem of turning insects into beasts of burden, and to develop technology for the domestication of the locomotion of insects,” says Bozkurt. “It’s a natural strategy to use the insects themselves.”

A Natural Strategy

The technique is simple enough. Get a sample of Gromphadorhina portentosa, aka, a Madagascar Hissing Cockroach, give it a circuit board for a saddle, then wire up the antennas. Using a joystick, wirelessly deliver pulses on the order of a few hundred millivolts. (Pulses, because continuous stimulation of the antenna can cause tissue damage.)

The roach, presumably, will think something’s in its path and move accordingly. Some do, anyway. Only about 1 in 10 of the cyborg roaches Bozkurt attempts to dominate successfully follow the path dictated by the joystick. And the obedient ones are only slightly more predictable than their wayward brethren. Their tendency is to wander randomly, and the joystick is used to steer them back onto the desired path.

Just how much they turn with each pulse has yet to be determined. The researchers did notice that the longer the pulse from the joystick, the larger the turn of the cockroach.

For the roach-rights activists, Bozkurt is quick to point out that technology that gets animals to go where we want them is not exactly new. “It can be a horse, oxen, a camel, or it can be an insect. You will need a technology to communicate with them,” he says. “A horse has an eye cover and reins. Normally, we don’t think of those things as technologies, but they are indeed technologies.”

As a graduate student, Bozkurt had once tried such technologies on the thornier problem of controlling the flight of a hawk moth, whose travels are more three-dimensional than the paths of surface crawlers. “When I became a professor, I decided to work on an easier problem,” he says.

Unpredictable Environment

Cockroaches also have the advantage of being nearly blind. Unlike bees, houseflies, and moths, all of whom have huge eyes that take up most of their head, roaches do the majority of their navigating with their antenna. Take control of those and you’ve taken control of the entire animal. Mostly.

“As engineers, we are used to working with systems that are generally predictable. Whenever you give an input, you generally get some output,” says Bozkurt. “Most of our systems are deterministic, but insects are affected by a lot of stimulus in their environment. We try to control them but they react with more than one sense. This is the major challenge we are trying to solve.” Another is shrinking the circuit, so the battery can get bigger and the insect can be sent on longer missions.

Once these problems are solved, or even partially solved, the roaches could be marched into places where no other animal or robot can go, such as search-and-rescue missions through the rubble after an earthquake, or data-gathering operations in a radiation-contaminated area.

Michael Abrams is an independent writer.

Our aim was to solve this problem of turning insects into beasts of burden.

Prof. Alper Bozkurt, North Carolina State University


October 2012

by Michael Abrams,