MIT’s Undersea Robots Can Plan Their Own Missions

By Carl Engelking | May 11, 2015 4:33 pm
A Slocum glider, used by the MIT team, navigates underwater. Credits Courtesy of the researchers

A Slocum glider, used by the MIT team, navigates underwater. (Courtesy of the researchers)

Engineers at the Massachusetts Institute of Technology have a plan to make ocean exploration easier: Let undersea robots make their own decisions.

Scientists use autonomous underwater vehicles (AUVs) to map the ocean floor, monitor the health of fisheries and study ocean habitats. However, writing computer codes that tell a robot how to accomplish each mission plan is incredibly time consuming at limits the capabilities of AUVs. But MIT’s Enterprise programming system gives AUVs “cognitive” capabilities, which allows them to figure out the best way to accomplish a mission without relying on an engineer’s list of commands.

Setting a Goal

With the new Enterprise system, humans can assign AUVs general goals, like, for example, exploring a series of locations in the ocean within a given time frame. Once it’s given a general objective, the AUV automatically crafts a plan to accomplish its mission in the most efficient way possible. If an unexpected event prevents an AUV from completing its mission while trolling the ocean, it draws up new plans and even fixes hardware malfunctions without human support.

MIT researchers tested the Enterprise system with an underwater glider in March during a research cruise off the coast of Australia. The robot adapted its mission plan on the fly and avoid collisions with other autonomous vehicles in the area, while still accomplishing its overall goal.

MIT engineers hand “cognitive” control to underwater robots Description Researchers watch underwater footage taken by various AUVs exploring Australia's Scott Reef. Credits Courtesy of the researchers

Researchers watch underwater footage taken by various AUVs exploring Australia’s Scott Reef. (Courtesy of the researchers)

Enterprise utilizes the same top-down command system from the “Star Trek” franchise, which MIT engineers modeled and named the system after. The first component of the system acts as the “captain” and oversees all facets of the mission. The second component functions as the “navigator,” planning out a route to meet mission goals. The third component works as the “engineer,” diagnosing and repairing problems. Throughout an entire mission, the “captain” continuously solicits feedback from these subordinates and uses the information to ensure it finishes the job.

MIT researchers will present the Enterprise mission-planning system in June at the International Conference on Automated Planning and Scheduling in Israel.

Exploring More

Researchers believe their Enterprise system will create opportunities to explore previously inaccessible nooks and crannies of the ocean floor. If AUVs can make their own decisions, they don’t need to rely on cable tethers that keep them in constant contact with humans. The technology, researchers say, will also free up mental bandwidth for oceanographers and engineers so they can focus on the big picture, rather than the minutiae of every mission.

The ocean remains the final unexplored frontier here on Earth, but mission-planning robots appear poised to lend a helping hand.

  • Uncle Al

    Let undersea robots make their own decisions. That is smarter than anything residing in the White House since 21 January 1989. Why not warm them? Youtube YQIMGV5vtd4

    • philconl

      They still set the mapping parameters, the area, the identification of the floor, the width of viewing map, turn radius parameters, etc.
      What they have found is the computer can be programmed to calculate a path faster than any of them can themselves, much like a calculator computes a large multiplication or a sine wave problem…

      It’s all programming in the beginning, and the end. Two teams clustering in on each others ocean floor don’t have to fight over who gets their robot schwimmer running view lanes first – collision avoidance takes care of that – if 10=hit, go to 20, 20 turn left 3 degrees, go to 10….

  • Adjoran

    Why no trigger warning for the giant phalli raping the ocean? Where is the safe space? And to have them “explore previously inaccessible nooks and crannies” is clearly microaggressive and represents the cis-heteronormative patriarchal oppression of wymyn.

  • Laccand

    < ++photos.essence++ see MORE DETAIL



  • Thoom1952

    < col Hiiiiiii Friends….< ++discovermagazine++ see MORE DETAIL

    ~~~~~~ ~~~~~~ ~~~~~~~~~ ~~~~~~~~~~~ ~~~~~~~~~ ~~~O

  • philconl

    “even avoiding other autonomous robots” – ROFLMAO
    The gigantic ocean and yep – their super smart AI human functioning “not programmed with commands” (WRONG !) water robo drones are running paths across each other because they are so “efficient”.
    They still set the mission scope and the loose parameters for accomplishing said mission, it’s ALL STRUCTURED COMMANDS EVEN IF LEEWAY WITHIN IS GIVEN.

    I am so sick of the spins and lies.



From studying wildlife to building drones that "think," unmanned vehicles are changing the ways scientists work and the questions they probe. Here at the Drone360 blog, we’re keeping tabs on the exciting ways drones are advancing science every day.

If you were a fan or subscriber to Drone360, this is your new home for all your rotor-relevant news. Welcome!

Buy the Magazine

drone 360 cover

See More

@Drone360Mag on Twitter


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Collapse bottom bar