Monterey Bay Aquarium Research Institute
Mission planning and day of operations.

"AUV Operations at MBARI" presented by D.R. Thompson. MTS/IEEE-OES Oceans2007 Vancouver conference

Mission planning

Mission plans for the upper-water-column vehicle are generated by an interactive GUI. The operator defines a box on an electronic chart, and the software constructs a "lawnmower pattern," which fills the box with waypoints and a survey line of AUV tracks that will allow the sonar returns to slightly overlap.

We plan missions manually for the upper-water-column vehicle. After several years of doing similar operations, we usually find ourselves making rather small modifications to previously planned missions. In the case of planning a new mission, we usually receive a series of waypoints from the chief scientist, or in some cases, a box to survey. We enter the waypoints into the electronic chart (WinFrog), and construct a survey line connecting the dots. Next, we ensure that there is a waypoint approximately every 3300 meters, which translates to about 40 minutes of running time at 1.5 meters per second, or 3 knots. When the AUV approaches a landmass or a sharp turn, we break up the preceding leg of the mission with another waypoint, so that the AUV will surface and get a GPS position fix about 1000 meters from the upcoming turn.

Our mission scripts have a set of global definitions at the top of the file. During the run, if the value of one of these items is exceeded, the mission will abort, and fifteen minutes later, the dropweight will release. The globals are: mission time in seconds, minimum depth (zero), maximum depth (275 meters, to avoid crushing the oceanographic instruments), abort depth (300 meters), abort altitude (2.5 meters above the bottom), and abort lockout depth (10 meters, to avoid false altitude aborts on schools of baitfish).

The mission scripts execute from the bottom up, but for clarity, we show a leg sequence top down. A mission leg is constructed from five behaviors, which execute in order.

A graph plotting the path the AUV will take around the Monterey Bay, including stop points to take water samples.
A graph plotting the path the AUV will take around the Monterey Bay, including stop points to take water samples.
Duane Thompson © MBARI 2007 1*

Day of operations


We prefer to prepare the mission script as far in advance as possible before running the AUV. This gives us time to check and re-check the script for errors and omissions.

The day before AUV operations, we verify that all instruments are installed and connected, put the syntactic foam blocks in place, and install the fairings. If any changes have been made to the vehicle configuration, such as adding or removing an instrument, we take the AUV to the saltwater test tank to insure the ballasting and trim are correct. We test the functionality of every instrument and subsystem. If the batteries need charging, we can let the charging program run overnight unattended.

We talk with the RV Zephyr crew to agree on a departure time and share an overview of the upcoming mission.

On the morning of the day of operations, we prepare the gulpers and install them into the AUV. Once again we test all instruments and subsystems. We make sure the Launch and Recovery System (LARS), capture pole, and recovery line are in good order. During the transit to the launch location we make sure the Iridium SBD messages are getting through, and continue to monitor the health of all the systems. If this is a newly scripted mission, we continue to examine the code until just before launch. When we arrive at the launch point, we turn off the deck power supply, and let the onboard batteries power the AUV. We ensure that all three batteries are discharging. Using the LARS lifting straps, we position the AUV on the deck cradle so that the nose will nest into the capture cushion, and then hoist the AUV into the upper stops. The captain extends the LARS out over the stern of the Zephyr, and down until the AUV is about a meter or so off the surface of the water. Just before releasing the straps, the captain surges the boat forward slightly. He watches the incoming swells, and times the release so that the AUV can exit the LARS without damage.

Once the AUV is floating in the water, we check the health of the system one more time, and await permission from the captain to start the mission. When we get approval, we start the mission by sending the mission script to the AUV from the control laptop across the Freewave ethernet link. We watch the syslog messages on the control laptop as the AUV goes through initialization of all instruments, gets GPS fixes, spins the propeller, turns into the correct direction, and dives. We generally stay in the area for a waypoint surfacing or two, to ensure that all is well, before starting other work, or returning to the harbor. Before we leave for the evening, we agree on tomorrow's departure time. We watch the SBD messages throughout the night on our home computers via the MBARI website.

MBARI staff put the external shell on top of the Gulper AUV onboard the research vessel Zephyr
MBARI staff put the external shell on top of the upper-water-column vehicle onboard the research vessel Zephyr
© MBARI 2010

Routine recovery

We have a good idea of the AUV's location throughout the night, and forecast its position for recovery. If the mission has terminated early due to a problem, we are able to predict an intercept position by plotting the drift.

We leave the harbor, and turn on the RDF unit. If it is still dark, we watch for the strobe. On the way out, we thread the capture line through a hole the LARS nose capture cushion and tie off the bitter end. We thread the wire portion of the line through a hoop on the end of a long extendable pole. Upon arrival at the pickup location, the captain extends the LARS until it's about a meter off the surface, and positions the Zephyr so that the AUV is just off the port stern. The ship's engineer mans the capture pole, and hooks the line into a latch on the AUV's nose.

Map of Monterey Canyon created with AUV data.the sediment transport waves seem to resemble glacier tracks.
Map of Monterey Canyon created with AUV data. The sediment transport waves seem to resemble glacier tracks.

We pull in the line and position the Zephyr so that the nose of the AUV nests into the LARS nose capture cushion, then pull hard to bring the rolling assembly all the way forward.

The captain reels up the lifting straps, and once the water has drained out of the AUV, brings the LARS back into the stowed position. We gripe down the LARS, set up the deck cradle blocks, and lower the AUV back down. Then we gripe down the AUV. Next we rinse the instruments with fresh water and install the deck tether. We turn on the deck power supply and take the load off the batteries. We download the data onto the control laptop, and check the recorded mission data for navigation, control, and scientific data.


Once the Zephyr returns to the harbor, we copy the data from the hard drive in the Main Vehicle Computer across the WiFi link to MBARI's Shore Side Data System (SSDS). The data is automatically processed by Matlab scripts into human-readable diagrams.

Ongoing and future enhancements

We always have several ongoing enhancements and improvements to both of the AUVs, such as installing newly requested instruments, installing doors in the fairings for less invasive access to accomplish frequent tasks, upgrading launch and recovery hardware, and redesigning electronics for greater efficiency.


1* Thompson, D.R.; , "AUV Operations at MBARI," OCEANS 2007 , vol., no., pp.1-6, Sept. 29 2007-Oct. 4 2007 doi: 10.1109/OCEANS.2007.4449212 URL:

Last updated: Aug. 17, 2010