Shape-Shifting Modular Sailing Robot "Protei Sputnik"




Introduction: Shape-Shifting Modular Sailing Robot "Protei Sputnik"

About: Education: - Sailing robots: Protei Inc & Scoutbots CEO Personal:

What is this?
Protei Mini-RC "Sputnik" is a remotely controlled (radio) modular sailing robot. The design is aimed at being easy to make in most makerspace, low cost, easily extendable. This is NOT a high-performance device and not fit for long-term navigation in the open ocean. This is a small prototype to get acquainted with Protei Shape-Shifting sailing technology, and a fun platform to experiment. If you already own a RC kit (receiver + transmitter) Protei MIni-RC "Sputnik" with 4 modules should definitely cost under 100 USD to build (in San Francisco).

How does it work?

Why doing this?
The ocean is where all life comes from, but it is dying. Acidification, plastic pollution, oil spills, radioactive spills, agriculture and industrial run-offs... Right now, to study the ocean, you would typically use satellite data our go with a big ship, with a team of qualified scientists, that is expensive, slow and dangerous. How can we make it affordable, fast and safe? Many universities and companies are working on developing surface vessels that would be effective and affordable, but no one is quite there yet.

This instructables is the continuation of Protei "Cardboard Model" built thanks to the support of Autodesk Pier 9 "Impact Residency Program" and the Autodesk Foundation.

Step 1: The Inspiration: Sputnik & Other Contemporary Ocean Robotics Devices

The main inspiration for this prototype is Sputnik.

Sputnik 1 (/ˈspʌtnɪk/; Russian: Спутник-1 [ˈsputnʲɪk] "Satellite-1" was the first artificial Earth satellite. The Soviet Union launched it into an elliptical low Earth orbit on 4 October 1957. It was a 58 cm (23 in) diameter polished metal sphere, with four external radio antennae to broadcast radio pulses. It was visible all around the Earth and its radio pulses were detectable. This surprise success precipitated the American Sputnik crisis and triggered the Space Race, a part of the larger Cold War. The launch ushered in new political, military, technological, and scientific developments. From Wikipedia.

What I love about Sputnik is its simplicity and ruggedness in design, and how it motivated a generation into space exploration. I hope the same kind excitement happens with the oceans.

Of course it is also important to look into contemporary ocean exploration devices and in these last years there have been some really interesting developments. Just a few references:

  1. Saildrone:
  2. Wave Glider:
  3. Datamaran:
  4. UBC Sailbot Ada:
  5. Pilothouse:
  6. Ardusailor:
  7. Green powered sailbot:
  8. Epsom College Sailing Robot:
  9. YunZhou Tech:

Also, because of the nature of Protei shape-shifting sailing robot, it is important to mention robotic snake research

  1. Sneel, by Gabriella Levine:
  2. SnakeRobot S2:
  3. Titanoboa:
  4. EE 125 Robot Snake:
  5. Kraken :

And these projects are also very inspirational to me and feature technology Protei can definitely learn from

  1. OpenROV:
  2. Fathom:
  3. Ibubble:
  4. Octanis:

Step 2: Materials and Tools

Here are the main materials

  1. 3mm Plywood to cut in a 24 x 32 in laser cutter
  2. Food Container (RubberMaid Brilliance)
  3. Main mast (48 in reflective Fiberglass rod)
  4. Mast reinforcement & Boom (Welding Aluminium rod)
  5. Bungee cord
  6. Rubber bands
  7. Stainless steel screws
  8. Waterproof Wood glue
  9. Stainless steel pins
  10. Marine varnish
  11. Spray paint
  12. Cable ties (Nylon)
  13. Steel for ballast
  14. Servo motors
  15. Battery Cable
  16. Battery holder
  17. Batteries (4 x AAA 1.5V)
  18. Receiver
  19. Transmitter

For the tools

  1. Laser Cutter
  2. Drill
  3. Utility knife
  4. Glue brush
  5. Metal saw
  6. Metal files
  7. Regular pliers

Step 3: How Is This Going to Work? General Design Idea

I tested some of my assumptions on Protei Cardboard Model (instructable), So I had a rough idea of where I was going. I wanted something simple with a

  1. single hinge mechanism
  2. module size based on internal minimal components (batteries, receiver, 2 servomotors)
  3. Flexible scale design to avoid gaps on the sides
  4. Easy to laser cut and assemble with cheap materials

Step 4: Waterproof Box for Electronics? Really Really?

I have been buying Pelican Case for a while, they are amazing but so expensive. For this project, I figured it might be overkill to use such military-grade dry case. I read a lot of testimonials online and found the Rubbermaid Brilliance set had good reviews. I bought and tested other ones, but this one is the best I feel.

I immerse 2 different sizes of food container with steel ballast for 3 hours, and they came out without a drop inside. Approved.

Step 5: How to Power This Modular Boat? Servo Motors

I am using the container "Rubbermaid Brilliance Small size". I couldnt find it online, so I modeled it here:

For the servomotor. I found it on on 3D warehouse, thanks luismi911:

Next, I proceeded to check if the servos would fit, and yes, 2 of them would be comfortable in this small box, sticking out.

I will use 2 standard servos for this.

  • Servo A would control the sail / wing
  • Servo B would connect to the next module, and "bend" the boat

Step 6: How Can This Boat Be Modular? Why Does It Matter?

Can you imagine how cool it would be if we could design boats in a similar fashion we design trains? You could carry so much more payload. Each module could have it's own experiment. A compartmented fail-safe system. To test the idea I have been building land-based Protei, I called "Windtrain" to test the idea. It does work up to a certain extent.

With 2 modules

With 5 modules, 3.5m long machine.

With 4 modules, 6.5m long machine.

With 2 modules and flexible joints. It's been harder to build that for the sea.

Step 7: What Mast and How to Hold the Sail?

If you are like me, the imperial system remains a mystery. To choose tubes, it took me some time to figure. Having an image of a drill bits set helped me. This table also was helpful.

== Update ==
I was going to go for a standard electric duct, but luckily found an awesome 1.99 USD thin and strong " 48 in. Reflective Rod Orange" made of fiberglass. Below is the rationale for the PVC tube - that I did not end up taking.
== End of Update ==

I want to stick to standards as much as possible. So I looked at what I could find in a local hardware store, that's affordable. Home Depot to the rescue. Plumbing section. $2.15 USD, awesome.

1/2 inch pipes are 21.33 mm in external diameter

And weights 0.16 lb/ft, that's 0.2381062kg/m, or 238g for a 1m long mast. Pretty good.

I will stick to a mast that everyone can find in their hardware store: the schedule 40, 1/2' PVC tube. Too big for my taste, but easy to find, lightweight, easy to hack and very durable .

Lighter and prettier with better performance, you can go for "3/8'' Rigid Aluminium Tubing", it's $3.72USD for 3ft :)

Step 8: How to Make It Flexible and Solid in Some Other Places? Kerf!

To get that flexible pattern, I started using the "Fish Scale Kerf". But I had many issues with a laser cutter not working well and broke many many attempts. I found out that even if the fish scale may be better for water flow, the simple straight lines are easier to make on most materials allowing great flexibility.

I started by using kerfs I found in these instructables:

If you want to check out more amazing ways nature can inspire designs (biomimicry) have your mind blown here:

Step 9: How to Make a Heavy Ballast to Stabilise the Boat? Keel Weight

To keep the sailing boat upright when the wind is blowing in the foil, one needs to add weight at the bottom. If you add too much, the boat will sink. If you add too little, the boat will "lay down" too easily. If your ballast is too bulky it might slow you down a lot. I cut the metal piece a little longer than I would and I will carry a metal saw during test if I need to shorten it down.

A smart friend of mine suggested to use diver's weights - will do in next prototype! I also found that they are not as expensive as I expected. Also fishing supply may be a smart option.

Also fishing

Photos by the legendary Prof Mc Indoe / Instructables user page.

Step 10: How to Make the Sail Rotate Smoothly? Bearings

I wanted my mast to rotate super smoothly so I decided to mount my mast on ball bearings. Turned out my 1/99 USD fiberglass reflective rod is 8mm internal diameter, just like the skateboard bearings. I found that I could buy 8 bearings under 10USD, enough supply for 4 modules Protei "Sputnik"! Bought these bearing off the counter, in the nearest skateshop I could find open the Saturday of Thanksgiving: DXL

Step 11: How to Make Custom Horns for My Servo Motors?

Futaba provides a page with the dimensions of their Servo Motors horns:

I assumed I was going to use this one I found on 3dWarehouse:

I had to scale it to the size referenced by Futaba.
I did not buy the official Futaba brand, but another standard size servo that works just fine.

Step 12: Now, Time to Build. Laser Cutting Time!

For laser cutting, I used scrap wood I found in the workshop. It's 3mm thick and the work area is 24 x 32 in.
I used an Epilog laser that is a 100W at Pier 9 Autodesk.

  • Speed : 50
  • Power : 50
  • Freq : 500 HZ

Step 13: Assembly of the Hull

The assembly of the hull is pretty self-explanatory. Don't forget the flexible "Scales" when you glue the sandwich. I used waterproof woodglue.

Step 14: Assembly of the Keel

The keel is a little trickier. Make sure you try to stack the pieces and that they all align and fit before you glue it.
In the end, I used rubberbands to "clamp" the keel vertical, to make sure my keel is well aligned with my mast.

Step 15: How Can My Sail Be Efficient and Solid?

Right now I don't have the time to look into which NACA profile I should take, so I am just going to just wing it (sorry). Craig O'Donnel did his homework and I should too, but I am going to trust him on this one and go for NACA 0018. Please teach me how to choose an appropriate NACA profile in the comments if you know.

This website provide robust data and simulations, that's going to be very useful when I do CFD with Autodesk software, just to see if I set my simulation correctly. I am aware I should do it the other way round :

  • 1. simulate it
  • 2. simulate again and develop a better design from the simulationS.
  • 3. Make it and
  • 4. Test it at sea... and back to step 1

Once I get better at simulation, that's what I want to do. Less empirical, more based on simulation.

Step 16: Assembly of All the Modules

I transformed my office table into a sort of mini production line. Here in the picture holding Protei Mini-RC "Sputnik" is Gabriella Levine - dear friend.

Step 17: Show Time!

So, here is the Protei Mini-RC "Sputnik" exhibited. It still needs some work and be tested at sea!

Step 18: Putting Some Skin on These Bones

I tested 4 different papers and 4 different glues.


  1. Tracing Paper
  2. Craft Paper
  3. Canson 250gr white
  4. Poster paper 200gr orange


  1. Wallpaper glue
  2. Photo mounting YES glue
  3. Gorilla wood glue
  4. Gorilla super glue

They all have their pros and their cons. For strength and simplicity I will go with Gorilla Super glue for contact with the wooden structure but I will also wet the thick paper with wall paper to make it softer and easy to form.

Step 19: Cut the Slack

With an exacto, cut the paper that is not supported by wooden structure.

Step 20: Reinforce and Waterproof

I used long strips of water dipped in wallpaper glue to reinforce my joints. I used wood glue to waterproof the hull filling up all the small holes.
Let dry overnight if possible.

Step 21: Testing and Observations

Photos and Video by Charlie Nordstrom (storyteller83), Pier 9, Autodesk. You must see his "tear-away goat pants" - really.

Testing is a key step in the process. Nothing ever happens the way you wanted to. I made A LOT of mistakes that you can see in this test video.

What went wrong?

  1. Sail: too high, too heavy. A wing sail is not very intuitive as one cannot see the wind "blowing" into the fabric.
  2. Mast: It would be a good idea to have the mast go down all the way to the bottom of the keel.
  3. Hull: too narrow, needs to be wider and more stable. Not buoyant enough.
  4. Flexible flaps: too fragile, they broke almost immediately
  5. Electronic box: leaky. Sea water + Electronics = :(
  6. Keel: should NOT hold any air or bubble. At low speed, a NACA profile may be a luxury. Instead having an uninterrupted surface may be a good idea.
  7. Ballast: should have used a modular system. I will never buy clearance cable ties, these were UV damaged and I must have used about 50 extra of them as they kept breaking on me.
  8. Fabrication time: too long. I need to find a way to build boats faster
  9. Simulation: I would love to have a workflow that integrates computer-based simulations.

What was good?

  1. Food container: I really like this food container system. It is very waterproof. I will abstain from drilling in it next time.
  2. Bugee cords to hold the boxes: That was awesome to be able to change the content of the module so easily.
  3. Modular Design: It can work if the sail is not too high (unstable). The boat with 4 modules can bend at 90 degrees and make very tight turns.
  4. Material: Cheap and robust (except
  5. Pretty: "Sputnik" was pretty before the test. It is a nice display piece, but pretty bad in the water
  6. Test Location: Aquatic Parc in San Francisco , is a great testing location. Small waves, some wind and a nice small protected beach, close to Pier9. I will come back!

In the next step, learning from all these mistakes, I have many design modifications.

Step 22: Autopsy and Recylce

In these 4 photos, you can see some of the ad-hoc modifications I did during test and some of the damage.

  • Hull: too much air was getting trapped in the keel, making the boat very unstable. I stripped away the skin on the keel, making it less hydrodynamic, but more stable.
  • Sail: The sail was way too high and require too much righting moment. I cut 2 sails in half.
  • Electronic box: The battery pack was too high and the electronics suffered a leakage. Next time, I will change the battery design so that the box closes more easily
  • Motors: You can see, after only a few hours of test, the screws are starting to rust with seawater. That is ok for this prototype, certainly not if you want to keep that boat running for a long time.

Step 23: Conclusion and Next Design

These are things that need to change:

  1. Mast
    Mast should be shorter, lighter, require less righting moment.
    Hollow high-grade aluminium could be good. Carbon fiber, inflatable, anything that is lighter :)
  2. Mast base
    The mast base should run through the hull and inside the keel. For such a small model, strength is key.
  3. Sail
    The sail should be lighter, and better waterproofed
  4. Hull
    What is the "inside of the hull" should not allow any water in. When water comes in, it radically changes the buoyancy
  5. Flexible flaps
    They look and feel nice, but in the rough environment that is the sea, they are way too fragile and snap too easily.
  6. Connectors
    The fact that they need to be bent everytime I connect them is a bit of a hassle - compared to the awesome bungees.
  7. Keel
    At low speed is a NACA profile really required? Maybe not. Maybe it is a better idea to have a continuous surface - no space between the modules.
  8. Ballast
    It would be smart to use industry standards, such as fishing or diving gear. Large batteries could work well in larger prototypes too, if lodged in a VERY waterproof case.
  9. Material
    For a small prototype, laser-cut plywood worked well. Still requires a few hours to dry and varnish. There must be a simpler, cheaper and faster way to work.
  10. Test protocol
    I wish I brought my wind sensor and measured wave heigh and current speed. No matter how innocent the test is, it is important to collect environmental data so people who look at the video also understand the conditions of the test.
  11. Simulation
    I am very eager to learn about computer simulations, in particular fluid--dynamics to improve designs, save materials and testing time.

I am now looking forward to work on the next design from the learning of this prototype. Please share your feedback and advice how to improve the Protei technology of a shape-shifting, modular sailing robot. Thank you!

Epilog Contest 8

Participated in the
Epilog Contest 8

Be the First to Share


    • Mason Jar Speed Challenge

      Mason Jar Speed Challenge
    • Bikes Challenge

      Bikes Challenge
    • Remix Contest

      Remix Contest

    2 Discussions


    2 years ago

    I really enjoyed reading this, awesome! I would suggest you study the footy class rc sailboat, lots of good ideas there. For the hull consider basic laser cut frames covered with balsa wood. Alternatively vacuum forming is quite nice and simple. The sail could also be thin balsa over a lightweight frame. Your keel is much too big relative to the sail. Water is much denser than air so the area can be much smaller. I think for the footy class people often use model helicopter blades. Finally, forget about simulation. It is extremely difficult to get good results, and you have plenty of problems to solve before the potential performance improvements are going to make a difference.


    3 years ago

    This is really extremely cool. I can't get over how cool this is.

    It does, however, raise some questions. The biggest question I have is, if the idea is to make lots of little, autonomous sailboats for ocean science, doesn't that increase the amount of trash added to the oceans?

    I mean, eventually these boats have to fail, and with larger number of boats, you'll have more failures. Compared to a full sized boat, which also have a much smaller chance of failure (or trash generation) per unit, I feel like we're increasing the impact to the ocean, while trying to study it.

    Don't get me wrong though! I think this is a really cool idea that needs to be developed! It's like weather-balloon satellites and micro-satellites, in that it's opening exploration up to smaller institutions and individuals, and I think that's great!

    Furthermore, the modular design not only allows for a bunch of cool configurations, it also provides an interesting way of optimizing geometry to improve performance.

    I think this is really cool, and I'm looking forward to seeing how your tests go!