Amphibian Robots

A walking, swimming salamander comes alive, inside a computer

Auke Ijspeert and Michael Arbib of the Brain Simulation Laboratory at the University of Southern California in Los Angeles have developed a walking, swimming salamander that comes alive, inside a computer.

Understanding the complex behavior involved in switching from trotting to swimming could lead to a new generation of amphibious robots, say the researchers.

Ijspeert and Arbib wanted to investigate how behavior emerges from simple signals in a creature’s central nervous system. To accomplish this, they built a computer simulation of a salamander’s central nervous system, and superimposed it on a computer animation.

The resulting digital creature exists in a simulated world of flat ground and water. Everything from gravity to friction to inertia power (and sometimes hinder) it’s movements. “Being able to explore by crawling in and out of the sea is not a trivial problem,” says John Hallam of the artificial intelligence department at the University of Edinburgh in a report published by NewScientist. “There is a tremendous niche market for amphibious robots which could be used for navigation and exploration.”

What can be simple movements for a live creature can be all too complex for their robotic counterparts. For example, moving from water to land - or vice versa - is a tough problem for a robot, because it has to completely change its gait and adapt to the new environment, without stopping. Designers in the past tried to solve this problem by breaking down the problem into parts and solving them individually. Ijspeert feels that approach is too inflexible.

Living creatures cope and deal with this problem by using their sensory inputs as switches that turn different neural control mechanisms on or off. These - in turn - are transformed into complex and coordinated movements in the body. By studying salamanders, Ijspeert and Arbib were able to test various ideas about how different neural mechanisms worked - and see how vertebrates control their bodies.

By emulating natural oscillators in the brain that produce rhythmic signals for various types of movement, the researchers were able to produce quite complex behaviors. “The circuits are capable of generating trotting and swimming gaits,” says Ijspeert. As the robot “sees” water approaching, or feels it, a series of neurological switches make it change from the trotting oscillator to the undulating swimming oscillator.

According to Ijspeert, the salamander was an ideal choice because it has many similarities with humans, but on a simpler scale. “It’s a living fossil of one of the first vertebrates that made the transition onto land,” he says. He believes his research will help us learn more about our own control mechanisms.