Slashdot Mirror


Researchers Fly 50 Autonomous Planes Simultaneously

New submitter MagicRuB writes: Researchers at the Naval Postgraduate School in Monterey, CA recently flew fifty small autonomous planes together in what they claim is a "record-breaking drone swarm". These aircraft were built from lightweight foam wings with hobby-grade components, and were equipped with an autopilot running firmware based on the open-source Ardupilot project as well as a companion computer running custom autonomy software built on top of ROS and an 802.11n wireless device to communicate with other planes and ground stations. The researchers are using this swarm as a platform for advancing drone technology, and hope to see results implemented in agriculture, search and rescue, and defense applications.

39 comments

  1. FTFY by s.petry · · Score: 1

    >The researchers are using this swarm as a platform for advancing drone technology, and hope to see results implemented defense applications.

    "Swarms" are not needed in either Agriculture or Search and Rescue.

    --

    -The wise argue that there are few absolutes, the fool argues that there are no probabilities.

    1. Re:FTFY by JMJimmy · · Score: 3, Insightful

      But they are needed for Skynet!

    2. Re:FTFY by Anonymous Coward · · Score: 0

      Actually, since all the bees(*) are dying, drones for agricultural pollination might be a good thing to have as a backup.

      (*) Not all the bees.

    3. Re:FTFY by willy_me · · Score: 2

      No, but the technology used to make a "swarm" work is required if you want drones to operate alongside regular aircraft - ex, Search and Rescue. Remember the fires in California where the water bombers had to turn away because drones were taking photos of the fires? It was not due to a lack of airspace - it was due to the inability to guarantee that the drones would not impact the water bomber. "Swarm" technology could provide that guarantee.

    4. Re:FTFY by Anonymous Coward · · Score: 1

      Why not?

      People may still hold onto the idea of the slow-paced farm life, but agriculture is very time sensitive. You can't just set and forget your plants until harvest. It takes time to review them and having a fleet of robots to turn a morning's work into an hour task (or even do the work continuously) would be a huge boon.

      I hope it's obvious how search and rescue would benefit from more parallelization.

    5. Re:FTFY by postbigbang · · Score: 0

      I see a run on boxes of 12ga shells.

      --
      ---- Teach Peace. It's Cheaper Than War.
    6. Re:FTFY by ShanghaiBill · · Score: 1

      "Swarms" are not needed in either Agriculture or Search and Rescue.

      The Navy is not involved in agriculture, and S&R is not their primary mission. Drone swarms have military applications. They can be used to overwhelm an enemy's air defenses. Coordinated attacks by multiple drones can likely defeat manned air superiority fighters in the near future. The F35 will be obsolete before it is even fully deployed.

    7. Re:FTFY by Anonymous Coward · · Score: 0

      http://www.globalaginvesting.c...
      http://citeseerx.ist.psu.edu/v...

      Yep, no application for swarm behaviors in SAR or Agriculture. You're so right.

      I'm glad you're not MY senior system engineer / architect - you lack vision, and are more interested in writing a pithy punchline than you are in accuracy, which speaks volumes about your integrity and ability to see wider applications of new technology.

      Please, stay in your server closet for the rest of your career.

    8. Re:FTFY by Anonymous Coward · · Score: 0

      Good argument? I guess that's the way forward. Make an impassioned, well-supported argument, and then simultaneously (subtly) point out that your entire argument is founded on false premises.

      The bees are fine, there was a temporary drop in population levels for a couple of seasons but they have not only rebounded but in many areas are repopulating to a level higher than where they were before.

    9. Re:FTFY by Anonymous Coward · · Score: 0

      "Coordinated attacks by multiple drones" isn't particularly different from air superiority doctrine for the last 30 years, we just don't have to pay as many people to actually sit inside of them.

    10. Re:FTFY by s.petry · · Score: 1

      No, the swarm technology would not help with the CA fires or anything else you wish to point to outside of military application. What you may be intending is something similar to a person claiming that swarms will protect agriculture by shooting bugs with tiny lasers. Which makes for a science fiction plot, but has little to do with reality or even the near future.

      When every single aircraft is controlled centrally we can talk, but until then it's fantasy.

      --

      -The wise argue that there are few absolutes, the fool argues that there are no probabilities.

    11. Re:FTFY by Anonymous Coward · · Score: 0

      If I were in need of rescue, I'd rather have a swarm of drones looking for me than just one.

      Likewise, a swarm of drones can e.g. spray a field much faster than just one.

    12. Re:FTFY by s.petry · · Score: 0

      The first document describes uses for robots in agriculture, not swarms. Then we have the transformers being described in your second. Tell the truth, you have not given any thought to the topic and just found something convenient in a week attempt to argue and look intelligent.

      Or hell, maybe you ride around on your enchanted unicorn uncovering pots of gold at the end of all them rainbows..

      --

      -The wise argue that there are few absolutes, the fool argues that there are no probabilities.

    13. Re:FTFY by Anonymous Coward · · Score: 1

      No, what the GP is suggesting is that the implied ability to fly in close formation without collision would mean a future drone, so enabled, would be less likely to hit a regular aircraft and so drones could be used alongside regular aircraft and a group could be brought to the area as a swarm.

      Drones could be deployed to provide additional information, such as temperature sensing over a wide area to help direct firefighting efforts.

    14. Re:FTFY by Anonymous Coward · · Score: 0

      Ahem. From the first document:

      Dorhout has developed a robot he calls Prospero. Using game theory and swarm behavior algorithms, Dorhout predicts that future farming will be accomplished by swarms of Prosperos who will plant, tend and harvest crops by optimizing yield at every inch of a field. Prospero is an off-the-shelf, low-cost (about $8) robotic platform. Dorhout wrote an advanced walking program that enables Prospero to move in any direction, then added a sensor array, LEDs in its belly, a photo resistor, seed hopper and fertilizer sprayer apparatus and turned it loose.

      Swarm behavior allows for optimized yields via more precise placement of seeds, better coverage of fields, and less wasted fertilizer and pesticide. How? By allowing a bunch of small, low-cost robots to parallelize a task that would otherwise require a multi-million dollar combine and harvester, which will spray hundreds of gallons of wasted fertilizer and pesticide all over the fields, and still end up with a sub-optimal seed placement & depth. Yep, no use in agriculture at all.

      If you don't like the idea of swarms of robots self-assembling and adopting different shapes, here's a simpler example, which has actually been built, and uses swarm techniques:

      http://www2.isr.uc.pt/~ali/MAR... -- from the paper:

      The GUARDIANS robot swarm is designed to be deployed in a large warehouse in the event of a fire to assist fire fighters. The large size of the environment together with development of dense smoke that drastically reduces visibility, represent major challenges in search and rescue operations. We discuss a swarm of robots assisting a fire fighter. We explain the swarming algorithms which make the robots follow a human while no communication is required. Next we discuss our method for cooperative localization system which uses wireless network as a tool. We also discuss the interface between a human and the robot swarm. The swarming algorithms implement the functionality by which the robots react to the human. The robot swarm can provide guidance information. Together with the fire fighters we explored how to understand and benefit from a robot swarm. Finally the paper presents experimental results of our sensor based exploration method.

      Tell the truth, you have not given any thought to the topic and just found something convenient in a week attempt to argue and look intelligent.

      Back at ya, champ. Your inability to argue and, apparently, use Google, also makes me doubt the value of that "architect" title, too.

    15. Re:FTFY by s.petry · · Score: 1

      Dear Derp,

      When you get past the first paragraph and read the details, you will find that the uses are for "if a plant needs water a robot gets water to the plant", and "if the plant needs fertilizer a robot injects fertilizer." and "if the plant has fruit a robot can get the fruit based on criteria". None of those things require swarm behavior and some involve much more than our current drone could do even if swarming. What it does require is reading past the first paragraph to actually gain some knowledge.

      Save your sarcasm for when it's actually viable, or just be a dumb asshole. The choice yours of course.

      ps Let Starscream know that Reality said hello for us!

      pps No wonder you remain anonymous, then attempt to belittle people who are not.

      --

      -The wise argue that there are few absolutes, the fool argues that there are no probabilities.

    16. Re:FTFY by Anonymous Coward · · Score: 0

      The "swarm" here is just a few leaders with others slaved to them at different altitudes... it's still quite impressive logistically to get that many in the air, but the swarming here wasn't that impressive imho. Still kudos to the team it's quite a feat!

    17. Re:FTFY by Anonymous Coward · · Score: 0

      Swarms is just another word for real-world parallel processing. I can't imagine why you think real-world applications don't scale? It seems we do that all the time by putting more people on a task. Why would it be different if we use technology instead of people?

  2. Where's the video? by Anonymous Coward · · Score: 1

    If you're going to claim a record of 50 drones swarming. you get a fucking video of it. I saw like 3.

    1. Re:Where's the video? by Anonymous Coward · · Score: 0

      I saw like 3.

      No, no. They launched more than that. One at a time.
      Then you saw a few land. One at a time.

  3. they claim: record-breaking drone swarm by turkeydance · · Score: 1

    they needed: someone to advise them about unfriendly terminology.

  4. This Is Bullshit! by Anonymous Coward · · Score: 0

    A looped video of launch after launch after launch after launch? WTF?

    One does not simply claim to have simultaneously flown 50 autonomous aircraft without a video of a flying V slow roll, or at least a four point roll, down the runway!

    Do it right!

    1. Re:This Is Bullshit! by hardtofindanick · · Score: 1

      Yes this is indeed very disappointing considering the possibilities one can do with a proper swarm behavior:

      https://www.youtube.com/watch?...

  5. being simultaneous is for cows by Anonymous Coward · · Score: 0

    You are all simultaneous cows, the same time!!! You say MOOOO simultaneously, as if a million cows cried mooo and were suddenly told to stop, simultaneously. YOU SIMULTANEOUS COWS!!!

  6. waste by Anonymous Coward · · Score: 0

    What a waste, just to do it i guess...

  7. No AFRICANS present, I see... by Anonymous Coward · · Score: 0

    I wonder why that is? Lack of IQ, perchance? Say it isn't so!

    1. Re:No AFRICANS present, I see... by Anonymous Coward · · Score: 0

      All I see are AMERICANS.

  8. Zerg rush FTW by Anonymous Coward · · Score: 0

    Nice to see life imitating art.

  9. Fly or stand by? by QuietLagoon · · Score: 1

    Did the researchers "fly" 50 autonomous planes, or did they just stand by while the planes flew themselves? If the researchers did "fly" the planes, why are the planes called autonomous?

    1. Re:Fly or stand by? by fhage · · Score: 4, Informative
      Find more details here The planes used a Linux based autopilot to control the aircraft which had waypoints sent to it via onboard Odroid SBCs. Primary communication to and from the planes used UDP packets sent through USB WiFi adaptors hung off the Odroids

      2 people "operated" the swarm, one commanding, the other monitoring. Planes got their swarm "slot" assignments at launch and tried to maintain position relative to a lead aircraft once they had reached their slot altitude. It's an interesting read. A quick estimate suggests they put around $1k in hardware into each plane. They describe guiding planes manually via spotter pilots using a secondary 900MHz spread spectrum radio link when anything went wrong.

      All planes were programmed with the same landing point, with the assumption that GPS and barometer inaccuracy would provide sufficient spread on landing. But with nearly 50 planes on the runway, on-deck collisions were unavoidable. Some of the video captured by GoPro cameras mounted to the nose of each plane show skidding into other planes, or coming down directly on top of other planes on the ground. Lesson learned for next time: add a bit more variation to the landing coordinates.

  10. Not autonomous by FrodoOfTheShire · · Score: 1

    If the researchers fly 50 planes then they are not autonomous. Better to say 50 autonomous planes fly while researchers look on.

    1. Re:Not autonomous by Anonymous Coward · · Score: 1

      If you interpret "autonomous" as equivalent to "decided by themselves to go for a little flight" then you are correct. Otherwise you are making a dumb point.

  11. Slow on the uptake? by felixrising · · Score: 1

    First posted about on the 3rd... http://diydrones.com/profiles/...

  12. Amazing! by Anonymous Coward · · Score: 0

    This is incredible! This means that AI is right around the corner. We are already half way there with Siri, and this was the missing piece of the puzzle. Too bad they weren't 3d printed or cloud enabled though.

  13. Planes on unique trajectories are easy by Anonymous Coward · · Score: 0

    Try quadcopters...
    https://www.youtube.com/watch?... (mind that was 2013).

    Really, it's more about setup and teardown, cause plugging/calibration/unplugging 50 units is a bear.

  14. Suarez... by Anonymous Coward · · Score: 0

    Kill Decision

  15. Naval Postgraduate School by Max_W · · Score: 1

    It is a military organization. The project is interesting, but I think any use of robots and AI for military purposes should be banned by a global UN treaty.

  16. shotgun... by Anonymous Coward · · Score: 0

    boom!!!1

  17. This Is The Shit! by Anonymous Coward · · Score: 0

    This is how to run an autonomous swarm!

    Too bad that needless drone paranoia and FAA bullshit is putting teh U.S. so far behind in this arena.