This month scientists published rare footage of one of the Arctic’s most
elusive sharks. The findings demonstrate that, even with many
technological advances in recent years, it remains a challenging task to
document marine life up close.
But MIT computer scientists believe they have a possible solution: using
In a paper out today, a team from MIT’s Computer Science and Artificial
Intelligence Laboratory (CSAIL) unveiled “SoFi,” a soft robotic fish
that can independently swim alongside real fish in the ocean.
During test dives in the Rainbow Reef in Fiji, SoFi swam at depths of
more than 50 feet for up to 40 minutes at once, nimbly handling currents
and taking high-resolution photos and videos using (what else?) a
Using its undulating tail and
a unique ability to control its own buoyancy, SoFi can swim in a
straight line, turn, or dive up or down. The team also used a
waterproofed Super Nintendo controller and developed a custom acoustic
communications system that enabled them to change SoFi’s speed and have
it make specific moves and turns.
“To our knowledge, this is the first robotic fish that can swim
untethered in three dimensions for extended periods of time,” says CSAIL
PhD candidate Robert Katzschmann, lead author of the new journal article
published today in Science Robotics. “We are excited about the
possibility of being able to use a system like this to get closer to
marine life than humans can get on their own.”
Katzschmann worked on the project and wrote the paper with CSAIL
director Daniela Rus, graduate student Joseph DelPreto and former
postdoc Robert MacCurdy, who is now an assistant professor at the
University of Colorado at Boulder.
How it works
Existing autonomous underwater vehicles (AUVs) have traditionally been
tethered to boats or powered by bulky and expensive propellers.
In contrast, SoFi has a much simpler and more lightweight setup, with a
single camera, a motor, and the same lithium polymer battery that’s
found in consumer smartphones. To make the robot swim, the motor pumps
water into two balloon-like chambers in the fish’s tail that operate
like a set of pistons in an engine. As one chamber expands, it bends and
flexes to one side; when the actuators push water to the other channel,
that one bends and flexes in the other direction.
These alternating actions create a side-to-side motion that mimics the
movement of a real fish. By changing its flow patterns, the hydraulic
system enables different tail maneuvers that result in a range of
swimming speeds, with an average speed of about half a body length per
“The authors show a number of technical achievements in fabrication,
powering, and water resistance that allow the robot to move underwater
without a tether,” says Cecilia Laschi, a professor of biorobotics at
the Sant'Anna School of Advanced Studies in Pisa, Italy. “A robot like
this can help explore the reef more closely than current robots, both
because it can get closer more safely for the reef and because it can be
better accepted by the marine species.”
The entire back half of the fish is made of silicone rubber and flexible
plastic, and several components are 3-D-printed, including the head,
which holds all of the electronics. To reduce the chance of water
leaking into the machinery, the team filled the head with a small amount
of baby oil, since it’s a fluid that will not compress from pressure
changes during dives.
Indeed, one of the team’s biggest challenges was to get SoFi to swim at
different depths. The robot has two fins on its side that adjust the
pitch of the fish for up and down diving. To adjust its position
vertically, the robot has an adjustable weight compartment and a
“buoyancy control unit” that can change its density by compressing and
Katzschmann says that the team developed SoFi with the goal of being as
nondisruptive as possible in its environment, from the minimal noise of
the motor to the ultrasonic emissions of the team’s communications
system, which sends commands using wavelengths of 30 to 36 kilohertz.
“The robot is capable of close observations and interactions with marine
life and appears to not be disturbing to real fish,” says Rus.
The project is part of a larger body of work at CSAIL focused on soft
robots, which have the potential to be safer, sturdier, and more nimble
than their hard-bodied counterparts. Soft robots are in many ways easier
to control than rigid robots, since researchers don’t have to worry
quite as much about having to avoid collisions.
“Collision avoidance often leads to inefficient motion, since the robot
has to settle for a collision-free trajectory,” says Rus, the Andrew and
Erna Viterbi Professor of Electrical Engineering and Computer Science at
MIT. “In contrast, a soft robot is not just more likely to survive a
collision, but could use it as information to inform a more efficient
motion plan next time around.”
As next steps the team will be working on several improvements on SoFi.
Katzschmann plans to increase the fish’s speed by improving the pump
system and tweaking the design of its body and tail.
He says that they also plan to soon use the on-board camera to enable
SoFi to automatically follow real fish, and to build additional SoFis
for biologists to study how fish respond to different changes in their
“We view SoFi as a first step toward developing almost an underwater
observatory of sorts,” says Rus. “It has the potential to be a new type
of tool for ocean exploration and to open up new avenues for uncovering
the mysteries of marine life.”
This project was supported by the National Science Foundation.