The University Record, February 11, 1997

Cockroach 'biobots' may help scientists learn about environments

Former engineering student Todd Conklin (right) places a cockroach at the beginning of an experimental test track while Stephanie Grace Lin, Ann Arbor News photographer, observes. Conklin graduated from the U-M in May 1996, but returned to Ann Arbor several weeks ago to repeat earlier experiments involving mild electrical stimulation of cockroach neuromuscular systems.

Photo by Bob Kalmbach

By Sally Pobojewski
News and Information Services


We can learn a lot from cockroaches, according to a team of research scientists at the U-M. In one of several studies under way at universities in the United States and Japan, scientists have used mild electrical stimulation of cockroach neuromuscular systems to direct cockroaches to perform simple tasks. In these experiments, they eventually hope to learn:


Whether cockroaches, equipped with microelectronic sensing instruments, could be used as "biobots," recovering environmental information from their own natural habitats or from areas not readily or safely accessible to humans---such as underground pipes, toxic waste sites or nuclear reactors.


How cockroaches transport objects bigger and heavier than they are.


How cockroaches move so efficiently over rough terrain.

The experiments were conducted during the summer of 1994 by Thomas E. Moore, professor of biology and curator of insects in the Zoology Museum, and Selden B. Crary, assistant research scientist in electrical engineering and computer science.

Another key member of the research team was undergraduate student Todd Conklin, who graduated last May. Conklin's job was to implant tiny stimulus-wires in the vicinity of peripheral nerves located near the base of the roach's antennae and attach a microcable connector to the roach's back with warm beeswax.

Once the wires, connector and microcable were in position, the roach was placed on an experimental track. When Conklin transmitted a mild electrical signal to the stimulus-wires, the roach moved left or right and could be steered to walk in a circle. Roaches also were steered up an incline to collect hoops and drag them back to the starting point.

Depending on the animal, the frequency of the signal, and the positioning of the stimulus-wires, the researchers achieved a limited, but repeatable, degree of responsiveness to their "steering" commands. Researchers achieved "steerability" in about 30 percent of insects tested in the original 1994 experiment.

"The pulsed, low-power electrical stimulus produces only a mild tingling sensation to the tongue of a human and does not appear to cause any discomfort to the roach," Moore says. After about 15 minutes or when the animal no longer responded to the stimulus, it was returned to its cage. "Most roaches used in the 1994 experiment lived two to three years more in the laboratory without any evident ill-effects before dying of old age," he adds.

Moore and Crary emphasized that their experiment was extremely preliminary and that much additional research is needed before scientists will be able to determine the feasibility of biorobotic technology.

"Our work was just a starting point, but it does demonstrate that it is possible to use electronic stimuli to direct the animal's movement and learn more about cockroach locomotion," Moore says. "Non-cabled sensory packages may in the future help us learn more about how cockroaches interact in their natural environment without directly observing them."

"These animals have a tremendous amount of potential as a research model in robotics," Crary adds. "It would cost hundreds of thousands of dollars to build a robot that can do what these roaches can do."

Among other institutions conducting related studies are the University of Tokyo; the University of California, Berkeley; Carnegie-Mellon University; University of Illinois; and Ohio State University.