MBARI uses several types of robots for oceanographic research, each
designed with different strengths to accomplish different missions.
We put cameras, sonars, and other sensors on the robots, then send them
where it is too risky, too expensive, or the job is too repetitive and
boring for humans. These robots are really an enabling technology
because they allow our scientists to ask and answer questions that they
otherwise could not.
For detailed investigations over small areas and short periods,
remotely-operated vehicles (ROVs) are MBARI’s workhorse robots. They are
powerful and flexible–carrying high-definition cameras and modular
tool sleds that can be configured for a wide variety of tasks, from
collecting sediment cores to sampling hydrothermal vents to scanning a
larvacean with a laser. An ROV essentially gives us eyes and hands deep
underwater, letting us observe the environment and interact with it to
conduct our science experiments. Since it is connected to a ship by a
tether, we have high-bandwidth communication for video and sensor data,
and can keep a human pilot in the loop to drive the ROV and use its
manipulators to achieve complex tasks. The tether connection also gives
our ROVs plenty of power for light, sensors, and precise maneuvering
control. But that comes at a cost–since the ROV is connected to a
ship it has limited mobility, and all operations include the cost of the
ship, its fuel, and its crew.
For longer missions over larger areas, we use untethered robots. These
Autonomous underwater vehicles (AUVs) have more mobility, but limited
power and intelligence. They need to operate on their own, so they carry
batteries for energy, acoustic and radio transponders for navigation,
and complex software for mission execution, decision making, and fault
tolerance. Without a human in the loop, the robot must react on its own.
But it also is not subjected to fatigue or limited attention.
These vehicles are well-suited to long, repetitive, and well-defined
tasks like making a detailed bathymetric map, or yo-yoing down and back
up through the water column between specified waypoints, measuring
temperature, salinity, and chlorophyll fluorescence. We use large
propeller-driven AUVs for day-long missions with high-power payloads
(seafloor mapping) or fast missions with high-volume payloads (plankton
gulper) and small buoyancy-driven gliders with low-power sensors for
long-term monitoring missions lasting over a month. In the space
between, we have long-range AUVs that can operate several days to weeks
carrying low- and medium-power sensors, using both buoyancy engine and
Some of our investigations are focused on the surface–the interface
between the atmosphere above and the ocean below. Autonomous surface
vehicles, like the wave-glider, are ideal for these missions because they
can put sensors right at the boundary, above, and below. We can
communicate with our ASVs via radio or satellite, and they can navigate
using GPS, so they are easier to work with in some ways than AUVs. But
they also must deal with the challenges presented by other ships, waves, and weather. We are beginning to use ASVs for many of the well-defined, repetitive tasks
that ocean scientists used to need ships for. And at the same time we
are developing ways to use them as communication gateways and navigation
aids for submerged AUVs.
Together, all of these robots are really a set of tools to help
scientists study our ocean. To test a hypothesis, you need to choose the
right tool for the job–and in some cases you need to make a new tool.
Sometimes that means a new kind of robot, and other times it means a new
method: using several robots together as a team, and leveraging the
strengths of each team member. That is what we are moving toward with
our robots–using gliders for long-term monitoring, bringing in LRAUVs
and wave gliders for more intensive studies, and using data from those
to direct short-term, targeted missions for the larger AUVs and the ROV
with the ship.