An Update on our student research project

Hi all,

It’s been a minute since I’ve posted. Considering that I was, I think, one of the early members of the discussion group, I thought I’d have had more to say by now, but, you know, university, things don’t move that fast. :slight_smile: And, there’s a snowstorm blowing through here today, so a good time to do this…

Quick review. Elmhurst University is a small liberals arts university in the suburbs outside of Chicago. Our also small Physics Department has been building a mechatronics curriculum to give our students training for careers in case they decide not to go on the grad school (or, I suppose, even if they do). One of the kinda unique things about our department is that undergrads have opportunities to do research work, on a small scale, that they’d typically see in grad school on a larger campus. We’re adding projects that would give those students interested in mechatronics a chance to build things using real life conditions. I’ve described the beginnings of this project in other posts, so I think I’ll leave it at that.

And so, here we are with no oceans nearby, but a really big lake, and I’ve been laying out the outlines of what we might try to accomplish, and I’ve been hoping that some of that would dovetail with some of the Bristlemouth work so that we might be able to help out there in exercising the standards in various ways. So far, I haven’t found the “magic key” that would get us a developer’s kit, but you know, hope springs eternal. :slight_smile: And if not, that’s okay. I understand that the kits are going to folks who are actually trying to make a living or have the resources to make a significant contribution, and I’m just trying to give some students some real life experience. It’s all good.

As I’m understanding the discussions I’ve seen or heard, one major focus is connecting passive sensors using Bristlemouth. I can see a huge need there; while we’ll certainly go there with our ROVs, I’m a bit more interested in seeing if Bristlemouth can be used to connect together and communicate with components on the ROV itself, so, thrusters, lights, cameras, sample collectors, photon torpedoes, etc etc. We can exercise that with very simple to build ROVs.

And so, our current status, such as it is…

Overall plan is to build a series of ROVs with increasing complexity and depth capabilities. I’d like to drop one to the deepest part of Lake Michigan, which is either 923 or 985 feet deep depending on who you want to believe. Granted, not the ocean, but still deeper than most small ROVs of reasonable cost can reach. And that is I guess one our secondary goals. Can we build something that can go relatively deep on a limited budget? There’s a 2nd leg to this project flow that might be looking at “ROV swarming”., and there are some interesting communications challenges there.

I’m currently out shopping for the parts to build our first simple ROV. This one is basically being done just for the experience of building it. No particular requirements other than being able to put it in water and not have the whole thing short out. I should have that one built and ready to go in about a month.

The 2nd ROV gets more interesting. While we’ll build it with a tether for power and control, we’ll pretty quickly move most of the ROV management to a Raspberry Pi we’ll be adding and programming. It will control all of the equipment on the ROV. And yes, I can use things like USB and even Ethernet to provide connectivity and communications, but here is where I thought we could experiment a bit with Bristlemouth. Maybe it’s not exactly where Bristlemouth is headed, which is fine. But that linkage from the Pi through some interface to get instructions to things like thrusters or a camera, let’s say, sounds pretty interesting to me. We should be working on this ROV by March although I’m sketching out the plans for it now. Students will be coming into the project at about that time as well, I think.

I’ve been doodling around an idea where we’d have a “smart buoy” like a Spotter with communications. It is attached by tether to a docking station that has an ROV parked. Rather than collecting data from a sensor, tho, we would load a “mission” into the ROV to go off and do something, and then bring back the results. Or, a small swarm of special purpose ROVs operating in a synchronized fashion. Outside of this wacky idea, for our projects, we probably don’t need something expensive like a Spotter. The rest of the developer’s kit could be pretty useful, tho.

Now I don’t have any illusions here. We’re a small school with limited resources. This is a project about seeing what we can build by doing most of our shopping at Home Depot and an online electronics store. But, if I can get the students thinking about the possibilities, and given them the chance to try, then they will have gotten some important experience that they can take with them. I don’t know that we’ll be breaking any new ground here, but I can close with a little story. This past summer, I was up in Door County, Wisconsin (not that far, really from that deep point in Lake Michigan). My wife and I were visiting a nature sanctuary there called The Ridges. Not a very big place, but with a dedicated group of volunteers. Awhile back, they decided to see if they could grow some native orchids there. Nobody knew anything about it; they have no professional scientists on staff. But they tried, and failed a lot, but they started learning things along the way and finally succeeded. Now, scientists from around the world come to visit this group of volunteers to learn what they figured out. So, who knows, maybe we’ll build something that turns out to be interesting. And if we can help Bristlemouth along the way, so much the better.

1 Like

Thanks for the project update, @Rongee !

What platform do you think you will start with for the first ROV?


Hi @zack_j

Could you explain a bit what you mean by “platform”? But in the meantime…

The 1st ROV will be constructed from Sched 40 PVC pipe, easily available at local stores. The thrusters will be “home brewed”, 3 of them. Electronics are all in the control unit on the surface with a tether running to the ROV for power and control.

The 2nd ROV, as sketched out right now, will start with Sched 40 PVC, and have a “2nd generation” with a frame built from aluminum pipe. Or we may go directly to the metal pipe. That one will have 6 thrusters, possibly modified bilge pumps. Depends on what I can find and what depth ratings I can find. This ROV will have a Raspberry Pi controlling most operations with instructions from a surface tether. And this is where we get into the territory of connectivity and communications to the components. You and others probably know all this, but the Pi has general I/O ports that can connect to other controller boards that can provide signal and power. I’ve come across a couple during preliminary reviews that say that they can provide enough power to run more powerful thrusters. We’ll see. And maybe that’s a fine way to do it. But I’m also interested in a single interface that can talk to all of the components as needed. The intent here is that we could potentially program the Pi to be semi or fully autonomous. We can also add other devices to this basic “frame” including lights, cameras, etc. Depending on the interest of the team at that point, we could replicate that Raspberry Pi controller on different ROVs, experimenting with different shapes (3D printed, for example) or sturdier frames that could handle sample collectors and the like. That might include ROVs that mimic biological movements, ie, fins, different configurations of thrusters, etc. And, then, could we take that Pi in multiple ROVs and have them work in concert, communicating through some sort of acoustic or visual mechanism. Basically, open ended at this point.

There are some robotics competitions out there that the students might want to get involved with as well. That will kind of be up to them as to whether we build an ROV with that in mind. These will have the same needs as the others, and so, offer the same learning opportunities.

Happy to answer any other questions you might have.


Hey @Rongee,

I’m very sorry to have missed this thread when you originally posted - great update! It’s awesome to see everything you’re doing to make this project happen.

I can definitely understand your desire to use Bristlemouth for the ROV-type modules you want to build – believe me, I can’t wait for that either. We’re still aiming to support higher power through the electronics which would enable things like lights and thrusters to be connected with BM, but those boards haven’t been designed yet. As you can imagine, there’s a large demand for increased (lower power) sensor support, so most of the internal engineering time has been focused on that.

Because I’m excited (and maybe a bit impatient) I’ve already designed some higher-power modules that will accept high-power BM once its ready, but for now just run off of the power supplied through the connector without any communication electronics.

Since day one, I think we’ve been aligned that the best way to build an underwater vehicle (especially if students will be working an it and modifying it over time) is to make all the subsystems modular. That way, each system can be almost like an independent project.

In the meantime, how far along are you with the overall frame design for your first vehicle? Perhaps we could discuss clever ways to make it easy to mount and reconfigure modules once they’re ready. It would be great to trade thoughts on how to do that with you.

I’m stoked to be chatting about all this with you here. Thanks again for the awesome update!


1 Like

Hi @estackpole Eric!

Hope you had a great holiday season.

The very first vehicle, which I’ve kind of taken to calling our “Minus 1” (thanks to the recent movie “Godzilla Minus 1”) will be a fairly quick build just to demonstrate very basic operation. I expect a box of parts to arrive for that shortly. Plastic pipe, simple clamp fittings, tether providing power and control. Shallow depth. As our students come on board, I’m seeing it more as a demonstration unit for them.

The 2nd vehicle. I’m debating in my head right now whether I want to build a plastic pipe version of this, or go straight to something like aluminum. The frame would be aluminum pipe, and right now my plan is to use a base platform of aluminum plate drilled and threaded in a grid like you and I have discussed. We’ve got a CNC mill at the school that I think we could program to do the work; I have to see what size limitations there might be. I could do it in my workshop as well, manually, but the thought of drilling and threading a grid of maybe 150 holes or so sounds, well, less than exciting. Size? Currently thinking of something in the range of 12 x 18 inches, but that remains to be seen yet. How to attach things to the rest of the frame, ie, thrusters, etc, is a good question. Could clamp them, but I’d like something a little more solid; maybe braze threaded flanges of some kind to the aluminum frame? We could, I suppose, use bar stock instead of pipe and just provide threaded holes all over the frame. If I did that, the whole assembly could be screwed together, allowing us to reconfigure the whole ROV as needed.

Since I don’t have a developer’s kit yet, it’s been interesting, and probably a good exercise, to see how we’d power and control components without BM. Most of the motor control boards available for arduino and Pi are lower power boards, although there are a few that run at higher power. A board that could run a 12-24v motor at variable speed based on software control would be pretty interesting. I haven’t found one yet, although I did come across a person on Hackaday that had either built a controller board or was working on one that would run a 12v motor and was accepting instructions using Ethernet (I think; I have to track that down again). Depending on the sensor, I suppose that those could be connected to the general I/O pins, but I haven’t looked at that very closely yet. In any event, it’ll probably be a good exercise for us here to build a version “pre-BM” to have something to compare against should we be fortunate enough to be selected to get a Dev kit.

Anyway, that’s what’s wandering around in my head right now; more than happy to hear any thoughts you might have as well. The things I don’t know far outnumber the things I know right now. :slight_smile:



One other thing, @estackpole ,

Any configuration that we build that is being “managed” by, say, a Rasp Pi on board, will need some kind of waterproof “box” to live in, and that box would need to have connectors to, probably, everything on the ROV that does something. I would imagine that waterproof containers exist at reasonable cost that could go to 1000’ feet, let’s say, but I wonder about all those connectors. I know it gets done because commercial rovers exist that have those capabilities, but how to do it in a student project could be interesting. For connectivity, it’d seem like we’d almost need something like the back of a stereo system, ie, a grid of connectors, in this case, all waterproof with the connection array itself attached in some waterproof fashion. I guess if we connected to every pin that we’d need and ran wiring from that, we could flood the enclosure with epoxy, maybe, to protect the connections from shorting, and water pressure? Any thoughts on that? This might be one of those obvious things that I don’t happen to know anything about. It wouldn’t be the first time. :slight_smile:


@Rongee I like the thought path you’re on. Building a first 'quick and dirty" ROV just to get things going and help wrap your mind around what areas to focus on sounds great, and the small 2nd ROV, potentially with a machined threaded hole pattern for attaching payloads and modules sounds like exactly what I would do as well. In some designs I’ve worked on, I have a hole pattern on the “roof” of the vehicle below some flotation, and I also have a hole pattern on the “floor” which acts sort of like a payload tray. I’ve assumed a 50mm-spaced square hole pattern, and as I chat with people, it sounds like those holes being sized for M6 screws is the most popular recommendation. M6 seems quite big for 50mm spacing, but this way we could support high mechanistical load payloads (for example line release mechanisms) and for minimally loaded payloads we could also potentially use plastic screws which don’t have the corrosion issues metal screws may for long duration soaks.

I am certainly an advocate of encapsulating the various modules in epoxy or something like that once they’re working so that they are independently waterproof and robust, and using some sort of universal waterproof connector to plug everything together. It would be great to get some Bristlemouth hardware into your hands to at least make mock-ups for this stuff, but that may have to wait just a bit longer…

As described previously, the current electronics boards we have for doing Bristlemouth communication aren’t designed to handle the power for many ROV things like lights and thrusters. Hopefully we’ll start pursuing options for that soon, but unfortunately we have to prioritize projects with low power needs right now with our limited inventory. Let’s certainly keep in touch as you’re getting ready to work on the second vehicle and maybe I can get some mechanical test hardware to you just to play with.

Here’s an ROV design I came up with some time ago that would use some of the prototype Bristlemouth modules I’m interested in building. It seems similar to what you were hoping to build so I’d love to hear your thoughts. Obviously these modules would be wired together with Bristlemouth Jumpers, but that was too much work to model, so this quick mock-up just shows the modules without wiring.

You can see there is a tilting observation platform (also with 50mm mounting holes) with a camera and light, and on the bottom is a mounting tray with batteries, and mounted on the sides and center bracket are thrusters. I’d build the whole thing out of plastic to reduce corrosion opportunities, and computation would be done for the most part in each individual module. The cool thing about Bristlemouth is that you don’t even necessarily need a central processer since each module is independently addressable. There are certainly advantages to having a dedicated central processing unit (which could also be a module), but theoretically, you could just command each module directly from the topside controller.

I think that building ROVs and other underwater systems in this way has tremendous potential, so I’m excited to keep exploring the possibilities together!

1 Like

Are most of the sensors built on metric sized frames, do you think?