Unmanned Sailboat Traverses The North Atlantic

Sailboats have been traversing the Atlantic Ocean since before 1592, sailing through sunshine, wind, and rain. The one thing that they’ve all had in common has been a captain to pilot the ship across this vast watery expanse, at least until now. A company called Offshore Sensing has sailed an unmanned vessel all the way from Canada to Ireland.

The ship, called the Sailbuoy, attempted the journey last year as well but only made it about halfway before the mission was abandoned. This year, however, the voyage was finally completed, and this craft is officially the first unmanned ship to cross the Atlantic Ocean. The journey took about 80 days using sails and a small set of solar panels to drive the control electronics.

Using this technology, the company can investigate wave activity in specific areas of the ocean without having to send out a manned vessel to install a permanent buoy. The sailbuoy simply uses its autonomy to stay in a particular patch of ocean. There have been other missions that the sailbuoy has been tasked with as well, such as investigating the aftermath of the Deepwater Horizon oil spill in the Gulf of Mexico. With a reliable craft like this, it becomes much easier, safer, and less expensive to explore the ocean’s surface.

Thanks to [Andy] for the tip!

21 thoughts on “Unmanned Sailboat Traverses The North Atlantic

    1. No market for it. It’s a solution looking for a problem.

      Say you’re a American or European company and you want to ship your goods across the Atlantic. Do you trust a purely robotic ship that can’t repair itself, deal with pirates or handle unplanned events?

      No. You want a competent crew on the cargo ship to make sure it reaches it’s destination.

      And then there is insurance. You think Lloyds of London is going to insure say $40 million in merchandise on a robotic ship? No!!!

      1. Why are you writing about cargo ships? This post is about a small vessel for completely different purposes.
        I also do not unerstand why removing a few humans from a multi million cargo vessel would be a big issue. It is not that their salaries have a big inpact on the operation.
        On the other hand, it removes the possibility of drunk captains driving their oil tanker into ice bergs or cliffs.

        And the Norwegians are already onto it.
        They are building a 120 TEU cargo ship with fully autonomous capability expected to be launched in 2019 and operational in 2020.

    2. Why do you think nobody tried it 20 years ago?
      And why do you think that they needed to do this 20 years ago?

      Just because something is possible, doesn’t mean you need to do it. Who’s going to fund it, what is the purpose of the craft and are there better/more reliable/cheaper/less complicated solutions to achieve the same goal?
      Measuring waves or sea conditions with a stationary buoy is much easier/reliable with a anchored buoy or measurement platform then using an autonomous vehicle. However over the past 20 years there have been huge improvements over computing power / GPS receiver technology and satellite communications options. Everything is much more energy efficient then it was 20 years ago. Therefore it is now easier to do these things with a much smaller setup of equipment. What was cutting edge technology 20 years ago, is a common building block today. There fore less batteries and solar cells are required, smaller antennas, larger data bandwidth for communication between the command center and the craft.
      So yeah… very much possible 20 years ago, but much more feasible to build and operate today.
      The question remains if the concept is an improvement over existing technology for marine activities (like data gathering and exploration). But does it really matter.

      Though what isn’t mentioned… but what really worries me! Is that these vessels can be made (or modified) to fit very small but profitable payload… if used for smuggling drugs… Being small means they are difficult to detect and by being cheap they can be deployed in huge numbers (by doing so, spreading the risk of failure/interception). By being autonomous it no longer requires a crew on the vessel to operate. I’m not saying that these boats will be used for this purpose, but I will not be surprised if they are…

    3. The last part of your comment is not true, many people tried it in the last 10-15 years. The micro transat challenge gather all the info about this: https://www.microtransat.org .
      From what I know of the field many things can go wrong during a transat. It is an endurance mission of several months in a very hostile environment: humidity, salinity, extreme weather conditions for such small boat. Plus the fact that it is not easy to learn from previous experience because when something wrong happen if you don’t recover the vessel (Which is super hard) there is very little you can understand of what went wrong. In the previous attempts several boats were lost because of fishermen (this can be visible from the GPS track), but many of them just stopped transmitting their position.

      The field of robotic sailing is becoming more and more important, a couple of companies are investing in it (saildrone.com for example), but if you look at the attempts of the micro transat challenge it is mainly hobbyist and university projects.

      Since we are on hackaday, I thought people could be interested in opensource sailing robot projects so here is a small list of teams I know that are doing it:
      University of plymouth: https://github.com/Plymouth-Sailboat
      Aberystwyth University: http://abersailbot.co.uk/
      University of Southampton: https://github.com/Maritime-Robotics-Student-Society/sailing-robot (where I am from)
      University in Shanghai: https://github.com/hywel1994/Sailboat-Ros
      Aland University: https://sailingrobots.ax
      Zhejiang university: https://github.com/zju-sailing-robot-team/sailing-robot
      UC Davis: https://github.com/ucdsailboat/autonomous-sailboat

      1. I am In the process off adding sailboat support to Ardupilot, would be great to get you opinion on how were doing so far. I looked at most of the teams above but was unable to find a nice list of features of each codes. I think building of Ardupilot we have the huge advantage of loads of supported hardware and software. Ardurover will shortly support fully autonomous sailing missions, I hope to merge in the next week or so! https://discuss.ardupilot.org/t/rover-sailboat-auto-mission/32650?u=iampete

        1. Oh, that’s really interesting. Yes, let’s keep in touch (you have our email address on our website: http://sotonsailrobot.org/ )

          Concerning our code (https://github.com/Maritime-Robotics-Student-Society/sailing-robot) the main “features” are:
          – 2 type of upwind tack-ticks: one based on laylines and one on a poll (=if the boat would be better on the other tack for more than 70% of the time over the last 10s we tack)
          – 4 different tacking methods (for different wind conditions) + an algo to select the best one (see https://blog.sotonsailrobot.org/articles/helming-procedure/ )
          – station keeping algo (method to stay within a defined radius of a WP)
          – (very simple) obstacle detection
          – and other things that I can’t remember now

          I am not super familiar with ardupilot/pixhawk but in your case when you program to go to a waypoint, all the computing is being done on the boat or do you regularly send back and forth commands (like when to tack etc.?)

          The only teams I know that are using a pixhawk are Zhejiang university (https://github.com/zju-sailing-robot-team/sailing-robot) and the Peruagus project (https://www.microtransat.org/2018_peruagus_boat.php not sailing boat, but still a nice project)

          1. Thanks, yes all computing is being done on board the pixhawk (there are loads of new cheap mini quad control boards that are now supported too). The features of the rover are so far:
            – Auto missions based on a heading controller for upwind and a l1 navigation controller for down wind, Supports tacking on maximum cross track error and tacking on geofence (more testing needed here but works most of the time)
            – Ardurover runs a rate based controller so tacking its limited to a maximum rate.
            – PID control heel, once a certain heel angle is reached the sails are released with pid control to maintain that heel angle.
            – Loiter round a waypoint.

            Using misiton planner we can also use spline waypoints and automatically generate mapping grids. Ardurover supports obstacle avoidance based on lidar and sonar, although this has not been tested with sailing but should be a easy job to add. In the future I’m hoping to add a optimization algorithm to automatically generate a polar diagram that can be used to select the best sailing angle for any desired heading.

            Would be great if you could try it out, should be a repetitively quick set-up. Although it does not yet support fancy magnetic encode windvanes, so far only 360deg potentiomiters have been implemented, but now its proven we can begin to add more driver types. Feel free join the discussion on on the Ardupilot forums. https://discuss.ardupilot.org/t/sailboat-support/32060/77

    1. maybe a semi submerged electric craft with solar panels just breaking the surface? Less sensitive to weather without sails and plenty of battery space for 24/7 operation at 2-3 knots.

  1. I seem to recall an autonomous diving drone being used several years ago to collect data about the gulf stream current. Used a greasy wax substance in a balast chamber to vary boyancy of probe. Probe would alternately dive to about 1500 feet deep then surface and broadcast readings to a satelitte relay. Power was obtained by a passive prop driven generator. Entire probe was powered by forward motion of alternate dive-surface cycle.

  2. Y’know, someone might want to talk to a sailor.

    It’s quite easy to rig a small boat to sail without any electricity involved; wind vane pilots have been employed for a couple centuries. There are many examples of unmanned/unpiloted objects managing transits of vast distances. A surfboard with a fixed cone sail of, mm, maybe 30 cm diameter, at the bow and launched from Newfoundland and Labrador will likely make it across if a little work is done to make it self-righting.

    It seems to me the problem is wanting to remotely control it. Or wanting a programmed course, rather than one chosen by the wind.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.