More Questions Than Answers
This is the beginning of my journey to build an unmanned surface vehicle (USV) that can autonomously navigate long distances in the ocean. I’m chronicling my experiences here in hopes that in the future this blog will help others who want to take on a similar project.
The project is currently in the very, very early stages, meaning that right now I have more questions than answers. However, I’m not completely starting from scratch, as I’ve already learned a lot from this project in which Damon McMillan (and his team) built a USV named SeaCharger, which navigated autonomously from California to New Zealand. Damon’s build-blog is fantastic and I highly recommend checking it out. I love SeaCharger’s simple and ocean-worthy boat design (see image below), and will most likely be modeling my own vehicle off of it.
Motivation. So why do this project? I’m taking on this project for a couple reasons. First, I want to gain more experience in several technical areas, namely autonomous vehicle control, perception and motion planning. However, in order to work in these areas I need a vehicle! There are of course simpler options than building your own vehicle, such as practicing with simulators and/or off the shelf robotics platforms, but surprisingly there aren’t that many options related to marine vehicles — hence why I’m building my own.
The second reason I’m building my own USV is that I love water environments. I love the idea of being able to spend time near lakes, rivers, and oceans while at the same time engaging my passions for software engineering, AI, and machine learning. Honestly, if I can use my technical skills to help explore, cleanup, and protect water environments then that would be a dream come true. Hopefully this project will lead me closer to that.
Finally, I’m a fairly outdoorsy person, and I love the challenges associated with using smart devices in harsh and remote environments. Using technology to explore and learn more about the hard to reach places, like the ocean and space, is just really cool to me.
The Game Plan. Ok, so those are the primary motivations, now how is all of this going to work? Excellent question! Here’s where I’m at. As of right now, the vehicle I’m planning to build will largely follow the same design as SeaCharger (see image above). Luckily Damon’s blog has some great advice on how to construct the boat. I’m really looking forward to building the foam mold and doing the fiberglassing! I’ll probably build a scaled down version (~1m) of the boat first just to accelerate the learning process.
In terms of electronics and motors for the boat, it should be similar to SeaCharger in that I’ll need:
- A motor and thruster to power the boat
- A motor and rudder to steer
- A bunch of batteries to power everything
- A compass to determine heading
- A GPS to determine location
- A satellite connection for receiving remote commands and transmitting data
- Solar panels to provide power
Where my design will likely differ from SeaCharger is in having:
- An underwater sonar for object detection and avoidance
- Possibly an above-water camera for object detection, tracking, and avoidance (like ships)
- A computer (likely Raspberry Pi) for controlling components as well as running perception and motion planning algorithms
Of course that’s all of the high level stuff. I still have loads of questions about the details, such as:
- How do I waterproof all of this stuff?
- Am I going to be able to meet my power needs with the solar panels?
- Do I need permits to deploy this thing in the ocean?
- …and a million more questions
And that’s what this blog is about. Answering those questions along the way, and writing about them here to help the next person who goes down this path.
