I’ve been mulling over a special-purpose vehicle design that’s closer to a quadcopter than an ROV and am curious to hear what you all think. In short, it would be designed around a heavy sensor that needs to sit still for 20 minutes or so to do its job and then move a short (~40 foot) distance. I’m picturing something essentially just like a quadcopter, whose main cargo is its battery, but in this case with the heavy sensor in the middle, sitting or three or four feet and surrounded by enough flotation to make the whole thing just a little bit negatively buoyant, with 4 (or 8?) vertical thrusters out on short arms. No batteries, it would be powered from the topside. When it’s time to move, it would lift off and fly like a quadcopter to its next scan location. Thanks to the new DVL from WaterLinked, I should be able to feed some good velocity info to the Pixhawk running the show. I guess what I’m wondering is if ArduSub can work in this way or would I have to use ArduCopter, and therefore adapt the output from the DVL into an optical-flow sort of message format?
That sounds cool! I don’t think you can achieve that with stock Ardusub yet. I’m also not sure how well ArduCopter is going to handle being underwater…
That said, this is a cool application and should actually be doable without too much work.
The in-the-works integration with the Waterlinked DVL should work fine with both copter(provided it is underwater) and sub.
I am working on a 4 thruster design that would probably do what you want and sounds similar to what you are thinking about. The difference is that I am trying to make it a tilt- thruster configuration with the orientation of the thrusters controlled by servos. I think I can do it with Ardusub. it will probably be a few months before I have it completed.
Cool idea! This sounds like a great application for something like this.
The most important thing to keep in mind with a quadcopter configuration is that they create translational forces differently than an ROV. A quadcopter in the air is always producing a lot of thrust to lift its own weight. To move around, you tilt, which redirects a small amount of that thrust into the translational direction. If you’re neutrally buoyant, however, you don’t need any thrust to hover, so there isn’t any thrust to redirect into translation. That means that a neutrally buoyant quadcopter vehicle would have to turn 90 degrees and point into the direction of travel. If you’re slightly negatively buoyant, then you’d have to turn less than that, but still a significant amount.
I think that will make it difficult to work with and difficult to mount directional sensors like the DVL. I think a good alternative is to have four lifting thrusters as well as four vectored thrusters for the translational movement. Or alternatively, you could make a tilt-thruster version like @johng mentioned so that the vehicle body can stay level all the time.