Autonomous transatlantic sailing boat project







Captain Kidd plans to have the electronics "sleeping" for most of the crossing, waking up a few times a day to check status and make adjustments. This will hopefully minimize power consumption and minimize the amount of time that powered-up electronics might be exposed to salt water. 


He gave up on using a servo-pendulum system, going for a direct drive vane setup instead. The old system had too many variables with the complexity of the paddle in the water on a heeling and turning boat, and might not have survived intact. The new system gives up the amplified power from that paddle, but gains a lot in simplicity. The vane will be on a vertical post near the stern, which will drive a gear near the rudder post. A second gear attached to the rudder will directly control the steering. This will require experiments with several different gear.



With a large keel bulb, Captain Kidd was a bit nervous about how much spare buoyancy the boat will have when loaded down with electronics and batteries, assuming that all the internal spaces that aren't foam-filled will eventually fill with salt water. He had considered extending the hull vertically by an inch or so but decided against that. Instead, the plan is to widen the hull above the original waterline by adding a long foam wedge on each side and fiber-glassing over it. It is hoped this will give a few more pounds of buoyancy and an extra two inches of beam. 



Captain Kidd needs motors for two systems: 1. the mainsail control, and 2. the direction setting of the wind vane. The sail motor will drive a worm gear at high leverage (allowing it to hold position securely with the motor off). It will need modest torque and speed is not an issue. The Captain decided to go with simple DC gear-motors instead of more exotic brushless motors, which should let him skip the electronic 'motor shield' and drive them directly with transistors connected to the main power supply. In theory, brushless motors should have some advantages if the compartment is exposed to seawater, but simplicity is preferred.



There are to be two or three eight-inch long solar panels, mounted at an angle on the starboard (south facing) side of the deck. Previously, the plan was to make the deck cover shaped like an inverted V. But now the deck will be level to mount the solar panels.



Captain Kidd was planning to launch May or June from Cape Cod, but the schedule was tight. First order of business is to prove the new wind vane steering setup, while work on the solar panel / battery / motor begins. As soon as the vane is capable of holding the five wind-relative points of sail, he'll hook up the motors to the vane and mainsail and have the arduino take control. Initially it will be hardcoded time intervals for testing, then taking commands by modem. With the compressed schedule, the Captain may skip the cellular modem stage and go directly for the satellite modem. 




1. the GPS Logger sub-system, including solar panel.
2. the "Navigation" Light sub-system, including solar panels.
3. the SPOT Tracker sub-system, including solar panels.
4. the Autopilot sub-system, including GPS, PICAXE, and Rudder Servo.
7. the PICAXE Autopilot Software.
8. the Ship-to-Shore transmitter.
9. Electronic Compass

This is the first system we have seen using PICAXE and ARDUINO microcontrollers - which makes a lot of sense. So too the electronic compass.



In a months-long autonomous ocean crossing, you would not want to rely on the processor remembering the state of the wind vane or sail control and never getting out of sync with reality. You need sensors that let your systems read the actual state directly, falling back to processor memory only if the sensors fail. 

One of the simplest sensors is a tiny, $2 Hall Effect sensor, which detects the presence or absence of a strong magnetic field. These can be attached to a surface and coated in epoxy, giving them a reasonable chance of surviving in a marine environment. The sensor is the tiny black rectangle, about the size of a typical LED. 


The wiring is simple. The left pin gets +5v power, the center pin goes to ground, and the right pin give the sensor readout. That third pin should be "pulled up" to 5v with a 100k ohm resistor to keep the signal from drifting. This hall effect sensor is a switch - it goes from high to low with no intermediate states and for that reason CK is thinking of using an analog versions that gradually scales the voltage as the magnet approaches; that would give a bit more flexibility in controlling it. 



Captain Kidd obtained a weather station wind instrument and converted it for use on Adventure. He added an electronic compass and assembled (soldering over 140 connections!) a "motor shield", an add-on board that interfaces the Arduino computer to electric motors.

Parts for the self-steering gear were ordered from

A "Triple axis accelerometer and magnetometer" was ordered from Captain Kidd then found the LSM303 chip while looking for an electronic compass, which he purchased for $25 US, because it also has tilt/acceleration detection built in. A design objective is to have redundant systems of different design, installed in parallel with primary sensors. The LSM303 uses a more sophisticated communication protocol, sending and receiving all data using just two pins. It looks like there's a good software library online that even does some of the trigonometry.

The electronic compass is needed to indicate which way the boat is facing, especially when currents or weak winds make the GPS direction misleading. (If a boat is facing north and moving at 1/2 knot, an east-setting current can make it look like it's heading NE or ENE.) Because the boat has no way of knowing the true wind direction, I'll have to measure the apparent wind direction and estimate the true wind direction from there. This is a crucial calculation, as the rest of the navigation software depends on a reasonably accurate wind direction for setting the steering vane properly.



For many years Captain Kidd has had a backyard weather station, but two months after installation, the wind direction sensor failed, and it was reporting nothing but north winds for the last two years. CK finally ordered a replacement wind instrument (direction and speed) and installed a few months into this project. The new one worked fine.

Most wind vanes work via a potentiometer. A potentiometer is basically a light dimmer switch - the circuit's resistance varies from low to high as the dial is turned. According to this documentation, the resistance of the potentiometer should vary from nearly zero to 20K ohms as the dial turns, and it has a 20 degree blind spot that reads zero. The system can turn with the softest breeze you can feel, it's surprisingly sensitive. 

CK soldered the three wires into leads that can plug into a circuit board. One was designated "common" which was wired to ground. One was called "direction excitation", and that was wired that to 5V. The green wire is used to measure resistance. When resistance is high, it matches the yellow excitation line's voltage, when it's low, it matches the red "common" line's voltage. The Arduino measures the 0-5V range, and returns a value between 0 and 1023 that scales linearly with the voltage. (1023 = 5V, 511 = 2.5V, etc.)

CK wired it into the Arduino, wrote a simple control program from scratch, and we expect was very pleased to find it worked exactly as expected. He assigned 0V to 10 degrees, 5V to 350 degrees, and interpolated the values in between. CK then set the zero and 5V extremes to read as 0 degrees. Note that this isn't compass direction, it's relative to the boat's bow, so losing 10 degrees on either side of "dead ahead" really doesn't matter, as sailboats can't do anything useful while facing directly to windward - except of course for rotary sails - but that's another story.


The wind vane can be set into six positions relative to the steering gear's paddle: straight upwind (unused), straight downwind, and 60 or 120 degrees off the wind to port or starboard. With three magnets on the wind vane's gear, and three sensors on the disc below it, we can read which position it's in:

If the wind vane is pointed straight ahead, the two forward sensors would detect magnets, while the aft sensor would not. Each of the six locking positions for the vane will produce a different on/off pattern in the sensors. (Conveniently, the positions 'all on' and 'all off' are not used, and they're the most likely ones for failure modes in a broken power or ground wire.) 

CK will add one more magnet/sensor pair in the drive gear that turns the wind vane, to indicate when it has reached its locked position. With four magnets, four sensors and six wires, he'll have real-time verification of the vane's actual position.

For the sail position, CK is thinking of doing some sort of a gear track - a linear toothed rail - driven by a worm gear. Moving it aft will tighten the sail for sailing close to the wind; moving it forward will ease the sail for off the wind or downwind runs. A magnet mounted on the gear track, with three stationary sensors, ought to provide full control for the two or three sail positions needed. 





Captain Kidd's robot sailboat has no collision avoidance system, thus is not COLREGs compliant, but it comes into a category (a gray-area) that to all intents exempts small craft. Whereas, our (BMS LTD) main concern for a larger ships is full compliance.






Captain Kidd's capsize Youtube





The Microtransat Challenge:
The 2013 Microtransat Map - the French École Navale boat on it's way

The Microtransat Challenge for competition details
Snoopy's Windmill boat 

Windmill Boat video
The 2012 Microtransat Map for Snoopy and the French

SPOT Track from the Mail Online - busy shipping lanes
UK Winds , Sailflow Winds , Met Office Rain & Wind

World Sea conditions, Temperatures & Sunshine or for expected wave heights
Bournmouth live web-cam - weather and waves a few miles west of Barton-on-Sea

The original Press Release
Video of Snoopy's launch on BBC TV News - Robin gets soaked 

Family video of Snoopy's launch
2012 Snoopy-Sloop history video

The underwater TV recce/winch boat used in 2009  "Rescue of Snoopy"
Robot Boat Sounds: past, present and future video September 2012 Text-To-Speech

Video of Snoopy Sloop for 2013
Snoopy-Sloop 2013 video with latest boat, the 2012 attempt, all boats since 2008
Video of Snoopy's Atlantic Attempt on 23rd March
Snoopy's Atlantic Attempt on 30th March
Video of Snoopy Sloop 9 in June 2013 on Bray Lake

Jasper Coppings article in the Sunday Telegraph
BBC News report of "crash landing"
Snoopy prepared for self-navigating Atlantic voyage Paul Miles, Maidenhead Advertiser 5th Sept 2012.




This website is Copyright © 2013 Bluebird Marine Systems Ltd.   The names Bluebird™, Bluefish™, Solar Navigator™,Blueplanet Ecostar BE3™, Utopia Tristar™ and the blue bird and fish in flight logos are trademarks. The color blue is an essential element of the marks. All other trademarks are hereby acknowledged.


  Bluebird trademark legend, blue bird in flight logoBluefish flying fish autonomous robotic ships navigation system