A Yale University researcher has developed a robot that uses a "biomimetic" sonar system based on tips from the animal kingdom. Electrical engineer Roman Kuc, director of the Intelligent Sensors Lab, says his "Rodolph" should be particularly useful for undersea chores.

"Researchers prefer vision because they think they understand vision," says Kuc. But this sonar is so sophisticated that it is capable of identifying a small rubber O-ring and telling heads from tails on a dime. Engineer John Leonard of the Massachusetts Institute of Technology says Kuc's robot is "a big step because it's adaptive," which means that, like an animal, it can bob around, positioning itself to get an optimal look at its target.

The robot, described in the August issue of the Journal of the Acoustical Society of America, deploys features found on both bats and dolphins. The "mouth," which sends out high-pitched, dolphinlike clicks, is flanked by two "ears" that rotate like a bat's to point toward the source of echoes.

Going by observations of how dolphins detect buried objects, Kuc positioned the transducers on a robotic arm, at a 45-degree angle and at a constant distance--in this case, 15 centimeters--from the object. Kuc keeps refining his creation. Recently, he says, "we have introduced a barn owl aspect" to the robot. Owls have one ear that points higher than the other. This helps them pinpoint the elevation of a sound.

Kuc says his robot should be able to detect production flaws in manufactured items. But its main value is expected to be displayed underwater--finding buried objects and locating things too far away for a camera to see. He plans to add a sonar-guided camera to the system.

"The number of true believers in sonar in robotics is actually fairly small," says Leonard. He is one of them--he would like to use Kuc's device to "perform experiments that mimic dolphin echolocation," for developing better ways of navigating and locating objects underwater.