Visit Your Local PBS Station PBS Home PBS Home Programs A-Z TV Schedules Watch Video Donate Shop PBS Search PBS
SAF Archives  search ask the scientists in the classroom cool science
scientists from previous shows
cool careers in science
ask the scientists

Photo of Jamie Anderson Jamie Anderson as seen on Natural Born Robots: Swim Like a Fish

Click on Jamie's photo to read a brief bio.

q Alan Alda asked about the fins on your tuna and you said they were a regular "engineering-type" design. Why didn't you design them like a tuna's real fins? (Question sent by Jackie)

A The robot tuna was designed as a testbed to study the input-output relationships between the swimming movements of the body and tail and performance such as speed and power consumption. The pectoral fins were needed only to give depth control, not as a part of the propulsion experiments with the tuna. Thus, we chose not to try to duplicate the complexity of the real tuna's pectoral fins. Instead, we substituted conventional diving planes that are geometrically similar to the real tuna's fins. The pectoral fins rotate to allow the tuna to dive and change its depth. Motors inside the tuna's hull rotate the fins to enable it to dive to the necessary depth for an experiment and as was shown in the show, they are designed to break away in case of collision with objects. Real tuna pectoral fins are much more complex--in addition to rotating up and down, the fins also sweep forward and backward depending on how much control action the fish needs to dive or maintain depth. Perhaps we'll attempt a design that captures this functionality in the next robot.

q How fast can your robot tuna swim? Is it as fast as a real tuna? (Question sent by Greg, Saint Paul's School)

A The robot tuna can swim 2.4 knots (1.2 meters per second) or approximately 0.6 body lengths per second. This top speed is limited by the capacity of the hydraulic system which moves the fish tail. Real tunas can probably swim circles around the robot since they routinely cruise at 1-2 body lengths per second and burst to 10-20 lengths per second depending on their size. Actual speeds of real tunas the size of the robot tuna are not precisely known as there are no captive animals of that size (2.4 meters) that can be studied.

q How does the robot tuna maintain its buoyancy at various depths? And what kind of gyros control the pitch, yaw, etc? (Question sent by William)

A Although often a concern for underwater vehicles, maintaining buoyancy is not a problem for the tuna robot. Since the robot was designed for shallow water missions at swimming pool depths, we are able to control buoyancy passively by carefully positioning the appropriate amount of ballast lead. A detailed log of all the weights and buoyancies of each element of the robot is kept in a spreadsheet program which computes where the center of gravity and center of buoyancy are located. By keeping the center of buoyancy over the center of gravity in the same vertical line, the tuna is balanced. We accomplish this with lead ballast inside the tuna's pressure hull. Similarly, pitch and roll of the tuna robot are passively controlled with appropriate placement of ballast lead to make the robot swim level and trim.

Although you are correct in assuming that the tuna robot has gyros aboard, they are not presently used to control the position of the robot. Control of heading or yaw of the robot has been accomplished with the use of a magnetic compass.

q Will that tuna be for sale? I want one! How much is it? Will there at least be a model? (Question send by James, 6th grade student, St. Pauls School, FL)

A I'm sorry to report that we are not presently offering the tuna robot for sale. At this time, it is a one of a kind and we expect to continue our experiments with it. As for a model, I encourage you to make your own tuna robot model. Our project began with, you guessed it--an actual tuna. We practiced making a mold with a mackerel, which is a much smaller but similar fish that you can probably find in your fish market. First you need to find a box that is slightly bigger than the fish, perhaps a sturdy shoe box. Then fill the box halfway with sand and place the fish so that the sand comes level with the centerline of the fish. Put some oil or petroleum jelly on the top of the fish and then cover with plaster. When the plaster is hard, turn the mold over and cut off the bottom of the box. Take out the sand, cover the fish and plaster joint line with oil and put more plaster to fill the box. When the plaster is cured, cut the box away and pry open the mold. You'll probably want to throw away the fish at this point as it will be quite smelly. Now you have a two-part mold of the fish that you can fill with any mold making material to make your model.

q Once you get the tuna robot out at sea, how will you keep track of it? And what will happen if a shark comes along and tries to bite it? (Question sent by Quoc, age 12, Blanchard, Oklahoma)

A The current version of the robot tuna is not designed to go to sea so we don't have any way to track it. The current version of the robot tuna uses a six degree of freedom inertial measurement unit to estimate its position. It contains three accelerometers and three gyros which can be used calculate position and velocity. Unfortunately, inertial instruments are not able to work over long periods of time without accumulating errors in the measurements. When the sea-going version is built, it will likely have a global positioning system that is able to receive satellite signals which allow it to calculate its position. The satellite signals do not travel through water so the robot would be required to surface to find out where it is. There are acoustic sensors that can be used underwater but these require that the robot operate in a previously surveyed area where there are acoustic beacons already installed.

If a shark were to bite the robot tuna I imagine it would have a big surprise and probably lose a few teeth. The robot's pressure hull is made of carbon fiber and epoxy with aluminum stiffeners. Although the rest of the tuna is designed only for swimming pool depths, the hull was designed to withstand the pressure of one hundred feet of water depth. Hopefully the robot tuna can outswim any curious or hungry sharks!

q How can the robot detect where it is under water? I know about the sensors that other robots have to tell it to stop before it bumps into things and when to go around obstacles, but I wanted to know if the same sensor system can be used underwater. Also, does the lack of light underwater affect the performance of such sensors? Thank you and good luck on perfecting your robot tuna. (Question sent by Jameel, Grade 11, Homestead High School)

A Currently the robot uses inertial instruments (three accelerometers and three gyros) to calculate its position underwater. If we build an ocean going version of the robot, we'll probably use a variety of sensors such as inertial instruments, a GPS (global positioning system) receiver and obstacle avoidance sonar. Although we currently don't have an obstacle avoidance sonar system on the robot tuna, there are many autonomous underwater vehicles which use such systems very effectively. An acoustic beam is sent out and the reflection of the signal off of objects in the environment is measured. The time that it takes to get the signal back is measured and since the speed of sound in water is known (although this varies with salinity and temperature), the robot can calculate the distance to objects. When the distance is within a threshold value, the robot will avoid the obstacle by turning or stopping. Acoustic sensors do not require any light so there is no problem operating them in the darkest depths of the ocean.


Scientific American Frontiers
Fall 1990 to Spring 2000
Sponsored by GTE Corporation,
now a part of Verizon Communications Inc.