Basic Stamp Kits
Books for Kids
Robots at Work
Courtesy of New Scientist Magazine
By Jonathan Beard
After long hours spent studying fish swimming at the New England Aquarium, engineers at
the Massachusetts Institute of Technology have built the first robot fish. They see it as
a step towards a robot that could one day swim off on its own to do undersea research.
Charlie the Testing Tank Tuna, as the robot has been christened, is an attempt to get
around one of the most serious limitations of today's underwater robots: battery
You simply can't put enough batteries on board an autonomous underwater vehicle (AUV) for
long-term missions like exploring the mid-Atlantic ridge for a couple of months,"
says David Barrett, a graduate student in ocean engineering who is developing the robotic
fish for his PhD thesis.
The solution lies in a better power source or a better propulsion system, Barrett says.
"And since no one wants to put a nuclear power plant on an AUV, we're going down the
propulsion route." As fish have the most efficient undersea propulsion system around,
the researchers decided to use that as the model.
Charlie is about 1.2 meters long and has 2843 parts, including 40 metal ribs and tendons,
a segmented backbone, a large tail and a Lycra skin. Like the bluefin tuna that Barrett
observed at the aquarium, it propels itself by bending its spine, creating an impulse that
swishes its tail to propel it along. The power comes from six small servo motors, linked
internally to its body by the tendons, which set its spine in motion.
"It is the first real-world working robot fish and it swims in a remarkably fish-like
manner," Barrett says. However, this first-generation robot's mobility is severely
restricted. It is attached to rails in the test tank and an external control computer.
Although the robot tuna's swimming already closely replicates that of a real fish, Barrett
recognizes that there is still room for improvement. "We need to learn to steer the
tuna, and find out what the most efficient swimming stroke is and continually improve the
algorithms that control the stroke." When that is done he intends to move on to a
fully autonomous robot, tethered to a controller, but otherwise free-swimming.
According to Barrett, the team has an ambitious goal. "Within five years, we hope to
have a fully autonomous UV that we could throw into Boston Harbor, tell to swim out the
mid-Atlantic ridge, take some measurements, and then swim home.
"If a robot fish like this gets lost on its way back from the ridge, it could be
Robolobster that will be sent out to find it. A joint project between MIT and Boston
University has built what is described as "a lobster-sized robot that will duplicate
the crustacean's keen chemical sensing abilities". About 23 centimeters long, powered
by 16 AA batteries and mounted atop two wheels and a plastic caster, the robot can swim
and turn almost like a real lobster.
Lobsters locate prey by following chemical trails in the sea, detecting tiny
concentrations of molecules with their sense organs. Robolobster currently has
conductivity sensors that enable it to find plumes of saltwater in a test tank, but the
scientists envisage models with chemical sensors that could locate an oil leak or