This Autonomous Boat Went from California to Hawaii and Beyond

Drones & Vehicles Robotics Science Technology
This Autonomous Boat Went from California to Hawaii and Beyond

I struggle to descend the stairs leading to the sandy beach at Half Moon Bay, California, awkwardly hefting the 60-pound solar-powered SeaCharger atop my shoulder. Amid numerous “what-the-heck-is-that?!” stares, I perform some last-minute checks of the propeller and rudder, then wade out knee-deep and push SeaCharger as hard as I can toward the oncoming waves. Moving at walking speed, the boat makes it through the first several waves without flipping. Relieved, I make my way back onto the beach, then turn and watch my two-and-a-half-year project plodding westward, gradually disappearing in the whitecaps. A watchful bystander approaches me and says he’s sorry I lost control of my boat and that he’s sure it’ll wash up on the beach. I assure him that the boat is on autopilot, going exactly where it’s supposed to go. “And where is that?” “Hawaii.” The look on his face is priceless.

My good friend JT Zemp assists with a trial launch. Photo by Jillaire McMillan
My good friend JT Zemp assists with a trial launch. Photo by Jillaire McMillan

Indeed, the idea of this tiny, homemade boat surviving 2,400 miles of open ocean to reach Hawaii seems foolishly unrealistic. I know that more than anyone. With help from friends, I built the eight-foot long, foam-and-fiberglass SeaCharger in my garage — not for money or competition, but simply as a challenge. And a challenge it was. What started out as a yearlong project turned into 30 months of mistakes, compromises, and do-overs. So for the next few hours, I spend my time worrying and fretting, glued to my phone, waiting for each telemetry report sent by SeaCharger’s satellite modem. When it becomes obvious that the boat is still on track, I get in my truck and drive home.

Touch and Go

For the next day or two, SeaCharger appears to be doing remarkably well. It’s windy off the California coast, and I can tell from the onboard attitude sensor that SeaCharger is heeling sharply. Yet the boat keeps moving west, slowly but surely, reporting in every two hours via satellite. Even at night the boat plods on with stored solar energy in its large lithium-iron phosphate battery pack. Then, after only two days, the boat fails to send its normal update. It’s not completely unheard-of — sometimes satellites don’t provide a strong enough signal — but it’s very rare. So I anxiously wait another two hours. Again, nothing. The boat is dead, and I know it. It’s never missed two updates in a row. I tell a friend that it must have sunk or been eaten by a shark. My friend says the boat will be fine. I find his reassurance unreasonable but comforting. Sure enough, two hours later, SeaCharger miraculously checks in again, and I breathe a sigh of relief.

Testing out the power system to ensure there will be no hiccups once the boat is alone on the open ocean.
Testing out the power system to ensure there will be no hiccups once the boat is alone on the open ocean.

Problems occur over and over again during the next few weeks — the motor controller stops and has to be reset, strong currents almost halt the boat’s progress, a cloudy day causes the boat’s batteries to run out of juice. Each time, the only information I have is that the boat has stopped moving. It has no weather sensors and minimal diagnostic sensors. Extra sensors would have been expensive and risky. Increased complexity means there are more things to break.

But without weather or other information, my imagination runs wild. I often assume one of the boat’s four waterproof electrical connectors has sprung a leak. I built them out of brass plumbing fittings, added O-ring grooves using a Sherline desktop mill, and potted the electrical contacts with 3M epoxy. The entire boat is an amalgam of hobby-grade/homemade components and professional-grade ones. Using purely homemade parts might jeopardize the boat’s reliability, but paying for purely professional-grade parts might jeopardize my marriage. Any time SeaCharger stumbles, I wonder if I went cheap one time too many.

IMG_20160524_173038158
The solar panels are relatively fragile, but because they are all plastic, they’re almost immune to corrosive seawater.

One of the biggest questions is if the boat’s solar panels will survive. They’re not the type you’d find on a house, as those can’t be immersed in salt water. Instead, they are made only from laminated layers of plastic, without the typical aluminum and glass components found in other panels. If placed on a house, these thin, semi-flexible panels might not withstand tree branches falling on them, but there are no trees on the ocean. Extra marine sealant is applied to the panels where the electrical wires exit, and the panels are hard-wired to the boat instead of using connectors. Two solar panels are used for redundancy, but there is no redundancy in any of the other components on the boat, mostly to keep costs down.

SeaCharger uses off-the-shelf electronics as much as possible. The brains of the boat are an Arduino Mega, an Adafruit GPS, a satellite modem from Rock Seven, a compass from Devantech, and a battery protection/charging circuit from AA Portable Power Corp. A typical brushless motor spins the propeller and an R/C servo turns the rudder. I don’t worry too much about the reliability of the electronics, but I do worry about the motor and servo. Water isn’t the problem: the motor transfers torque to the propeller through a magnetic coupling, so it stays perfectly dry. And the servo has its own custom enclosure with rubber shaft seals to keep water out. But the bigger issue is the time required to get from California to Hawaii: the motor will have to run almost nonstop for over a month, while the rudder servo will have to complete 2 to 3 million cycles.

A: Electronics Enclosure B: Devantech CMPS-11 tilt-compensated compass C: Seeeduino Mega D: Pololu Servo MUX Board E: R/C Receiver F: Adafruit GPS G: Breadboard H: Servo I: Rudder control arm (connects to servo) J: Rudder K: Solar Cells L: Boat deck M: Hull N: Carbon fiber shafts with embedded electronics wiring O: Voltage Regulator P: Battery controller and D/C regulator Q: Data logger R: Batteries S: Balancing Board T: R/C airplane brushless motor U: 4:1 belt drive reduction using GT2 timing belt V: Thruster W: Propeller X: Prop screw Y: Keel enclosure Z: Black delrin thruster pod endcap
A: Electronics Enclosure
B: Devantech CMPS-11 tilt-compensated compass
C: Seeeduino Mega
D: Pololu Servo MUX Board
E: R/C Receiver
F: Adafruit GPS
G: Breadboard
H: Servo
I: Rudder control arm (connects to servo)
J: Rudder
K: Solar Cells
L: Boat deck
M: Hull
N: Carbon fiber shafts with embedded electronics wiring
O: Voltage Regulator
P: Battery controller and D/C regulator
Q: Data logger
R: Batteries
S: Balancing Board
T: R/C airplane brushless motor
U: 4:1 belt drive reduction using GT2 timing belt
V: Thruster
W: Propeller
X: Prop screw
Y: Keel enclosure
Z: Black delrin thruster pod endcap
Click for larger version
Diagram by Hep Svadja

Smooth Sailing

In spite of my worrying, after 3 weeks at sea, the boat is not only still alive but is actually moving along at a very good clip. For the past two and half years, amid my children’s constant cries of “Dad, when is the boat going to be done?” I’ve motivated myself by visualizing the moment when I’m standing somewhere in Hawaii as SeaCharger appears in the distance, comes motoring into the harbor, and I pull it triumphantly out of the water. Now it looks like that might actually happen!

After 41 days at sea, the SeaCharger takes a well-deserved break at Mahukona Harbor before the next launch.
After 41 days at sea, the SeaCharger takes a well-deserved break at Mahukona Harbor before the next launch.

Three more weeks pass. I’m standing on the shore at Mahukona Harbor on the Big Island with my wife, parents, brother, and a reporter from the local newspaper. I catch the first glimpse of SeaCharger’s solar panels flashing in the setting sun as it approaches. This moment is not as triumphant as it is surreal. This is the same SeaCharger that left California 41 days and 2,413 miles ago, but the faded paint and clinging barnacles only hint at what it must have experienced — and survived — to get here.

As I suspected, the boat accumulated a number of barnacles that slowed it down.
As I suspected, the boat accumulated a number of barnacles that slowed it down.

Safely ashore, SeaCharger appears to be in remarkably good shape. I’ve been asked dozens of times what I would do with SeaCharger after it made it to Hawaii, but I had never seriously considered it. Now it’s time to decide. We could pack it in a giant crate and ship it home, but that option doesn’t appeal to our pocketbook nor our sense of adventure. The truth is, the goal of this project was to send a solar-powered boat across an ocean. Making it from California to Hawaii was certainly an epic voyage, but it wasn’t truly transoceanic. So I reprogram the boat to navigate across the Pacific towards New Zealand, a seemingly impossible 4,400 miles away, and launch it.

The California launch was cold, windy, and extremely nerve-wracking. But in Hawaii, there’s only a minor breeze blowing over a perfect white sand beach. I wade in, waist-deep, lower SeaCharger into the warm water, and it casually motors out to sea. There are no nerves or tension. I feel SeaCharger has already proven itself.

I march toward the water on a beautiful day to relaunch the boat from Hapuna Beach, Hawaii, on July 27, 2016. Photo by Jillaire McMillan
I march toward the water on a beautiful day to relaunch the boat from Hapuna Beach, Hawaii, on July 27, 2016. Photo by Jillaire McMillan

Supremely Seaworthy

As I write this, 90 days have passed since SeaCharger left Hawaii. It’s crossed both the equator and the International Date Line and has traveled over one quarter of the Earth’s circumference. On one or two occasions it came dangerously close to running aground on a small island, thanks to a combination of my poor planning and stormy weather. Every time SeaCharger nears an island, I research that island’s geography and history. I learned that the Maori name for New Zealand is Aotearoa, or “the land of the long, white cloud.” Gee, it sure would’ve been nice to know that before I sent my solar-powered boat there!

The Maori knew what they were talking about: there are lots of clouds around New Zealand. And as SeaCharger makes its approach, the wind and currents conspire against it, pushing it north as it tries to make headway south. After three months at sea, the boat is dramatically slower, likely carrying a large load of barnacles. At one point I give up on New Zealand and send a command to the boat’s satellite modem to turn it west in an attempt to make it to New Caledonia or Norfolk Island. But the incredibly friendly, enthusiastic New Zealanders who are preparing for SeaCharger’s arrival assure me that conditions will change. Sure enough, soon the winds calm and southward progress becomes possible again.

New Zealand is only 500 miles away now. I have no idea if SeaCharger will reach the shore alive. Whatever happens, I am both supremely pleased and completely baffled that it has come this far and lasted this long.

Perhaps the greatest surprise of the entire project has been the intense interest and unfailing support from hundreds of people all over the world. I would expect the New Zealand high school science teacher who leads an autonomous boat club at his school, and the budding engineer who grilled me on the technical details at Maker Faire, to be interested in this record-breaking boat. But it’s even more surprising and satisfying to have my vintage-VW-loving brother-in-law and my wife’s gardening, baking, homeschooling mom friend tracking it faithfully online. I’ve come to realize that this voyage has a universal appeal far beyond what I ever imagined when I began.

An immense thank you goes to JT Zemp, Troy Arbuckle, and Matt Stowell. Troy and Matt built the rudder actuator and helped with various electronics. JT was there from the beginning, helping with overall system architecture, fabrication, and testing.

Epilogue: Adrift

Damon reports that on November 18, after 155 days at sea, the rudder stopped responding. This time SeaCharger really is dead, but not before it traveled an impressive 6,480 nautical miles.

Discuss this article with the rest of the community on our Discord server!
Tagged
Damon McMillan

Damon McMillan is a mechanical engineer, hobbyist and father of four. He resides in Sunnyvale, CA.

View more articles by Damon McMillan

ADVERTISEMENT

Maker Faire Bay Area 2023 - Mare Island, CA

Escape to an island of imagination + innovation as Maker Faire Bay Area returns for its 15th iteration!

Buy Tickets today! SAVE 15% and lock-in your preferred date(s).

FEEDBACK