Can we trick out a BlueROV to match true commercial subsea equipment?

I have REALLY enjoyed watching the ecosystem of low-cost ROV companies that have sprung up around the BlueROV project. There is so much cool stuff available now – Topside power supplies from outland, DVL from Waterlinked, USBL positioning from Cerulean Sonar, 360 sonar from Blue, custom controller panels, robotic arms, fiber optic tether conversion from SeaView systems, side scan, scanning and even multibeam sonar systems from DeepVision, Tritech, and others… This stuff is priced so that’s within the grasp of a bootstrapped small business.

But here’s the reality check… Not too long ago, I did a search and rescue project alongside a big company running a really impressive inspection-class ROV. I didn’t get the details, but picture something about like this:

They were able to plug in the coordinates of a target we identified with side-scan sonar, drop the ROV in the water, and navigate the ROV directly to those coordinates using a moving map interface fed by topside acoustic positioning and DVL. As they got close to the waypoint, the scanning sonar picked up the target, and they used it to guide them gently in for a close look with cameras.

“Nope, that’s just a dog crate.” So up came the ROV, winched onto the boat’s deck, and then off to check the next target.

I’m not going to lie, it was really f-ing cool to watch them work.

So… Are we to the point where we can replicate that capability with a BlueROV and third party parts? It seems like all of the individual components exist now, but is it possible to get all of them working together reliably in a single control interface like the one pictured above?

And if it is possible, what would the total cost be for a completely tricked out, maximum capability, BlueROV Max-edition?

I personally believe you could buy all the parts for under $50K. Compared to $300K+ (maybe way more) for a commercial rig.

But I’m less confident that there is an existing solution to make all this stuff work together. I don’t know. What do yall think? Has anyone done this?


By the way, anyone curious about prices for commercial sonar and rov equipment on the second-hand market should definitely subscribe to the DeepSearch and SurveySearch monthly listings published by (Scroll down to the bottom).

Most of that is possible today with BR.
You can do stuff that only is possible with workclass systems.

Still sometimes it is to much fiddling to get it working for a professional user, where it is not time to get in depth in the software djungle being offshore.

This is a very interesting topic that comes up every now and then! My two cents:

Small ROVs and sensors are certainly increasing the range of tasks that we can achieve without using large vehicles, however a portion of this is down to the fact that historically, many tasks have been carried out with unnecessarily large vehicles because that’s all we had access to! There are, however, some levels of performance or capability that small vehicles simply aren’t suitable for.

A lot of this comes down to availability of power, and also required accuracy of sensors. You mention the use case of plugging in coordinates and navigating to a pre-set waypoint. You can indeed achieve this with a BlueROV and small USBL/DVL etc, however the challenge is defining how accurately you need to be able to navigate, and the resolution to which you need to identify an object. At this point, the sensors available for a BlueROV might not match the job’s requirements.

Really this all comes down to understanding the error budgets that you have to play with. Smaller sensors are becoming much more capable, however there’s still a pretty large gap between guidance/navigation and survey grade (particularly with heading sensor technology).

One thing that makes a big difference when it comes to user-friendliness is having software that’s able to fuse data from multiple sources, and provide the operator with semi-automated functions. A good example of this is EIVA’s Mobula package, which interfaces very nicely with combined nav packages like the Nucleus1000 to give genuine navigation outputs, rather than a simple stream of data.

1 Like

Thanks for the thoughtful reply!

I think that more than anything else, this is currently the limiting factor. I am not an expert, but from my understanding of the technology, you have to combine the data from all of these sensors (inertial, magnetic, usbl, dvl, probably depth and altitude also) then feed them into a Kalman filter (or similar), in order to calculate position and orientation estimate.

Ardupilot performs this calculation with the sensors available on a Pixhawk plus GPS, in order to provide orientation and position information. Are there any open source or DIY oriented projects to extend this functionality to subsea vehicles by incorporating data from DVL, USBL and other underwater sensors?

It would be incredible if you could just add USBL and DVL data streams as inputs to Ardupilot, and suddenly the moving map display in Qgroundcontrol starts working for underwater vehicles also.

Although not survey caliber as Rory discussed, but the team over at Cerulean Sonar have done a great job with their line of sensors. By leveraging their Cerulean Tracker program, the sonar sensors (DVL, USBL, GPS) are fused through a Kalman filter that then gets fed to Qgroundcontrol. Sensor Fusion - Cerulean Sonar Docs


To me, the answer to your question is a definite YES!
I’ve used the BROV2 with limited sensors to find and retrieve targets, with a vessel crane, in depths from 50-100m. The limiting factor is current speed, tether length, and operating depth - with the small scale of the ROV governing most comparisons with the “work class” systems you mention. Pilot familiarity and skill can also factor into it, but that may exist at all levels.
A DVL or Cerulean product + heavy configuration + tether power supply really unlocks robust capacity in the price range you mention.

All that said - there are obvious benefits to larger size and power vehicles in some situations. I’d make the analogy that at this point, the BROV2 is capable of being tricked out in the same way an old vehicle could be - it may require expertise to use, and may lose certain challenges to modern vehicles, but with a bit of luck and preperation you can get the job done!

1 Like

This is an interesting discussion and I generally agree with Anthony but hopefully some of these limitations will be reduced as products are developed. There are some that we might not be able to get away from - such as current speeds and power or inertia required to fight a current or maintain position.
What has been your experience operating the BlueROV in high currents? We find it can shoot off once deployed and it usually pulls a load of tether with it. The tether in the water column has to be adding to the drag. I found a brief examination of tether drag here. I don’t want to go into detail on it because it may be off topic but I calculate approx. 300N drag forces assuming 100 m of 8 mm tether in 2 m/s water current. The BlueROV has a max speed rating but I don’t think it accounts for the drag on the ROV or tether drag. For this thread, can anyone offer their experience with max currents the BlueROV can fight on battery? Is surface power supply the way to go for currents?

We operate offshore from a boat in water about 100-150 feet deep. We’ve used two methods to remediate tether drag. The first is to rig a weight on the tether about 20 feet from the ROV, similar to this. The weight sits on the bottom near your target, and you are able to maneuver around it with limited tether drag on the ROV.

The second method, which we use on short dives, is to flake all of the tether loosely in a plastic bin or bucket on the boat. During the dive, we allow the current to pull the tether from the bin with no resistance. The tether is swept away from the boat in a huge arc by the current, but as long as it is freely released at the surface, there is little drag felt by the ROV. When we run out of spare tether, everything comes tight on the ROV, and the dive is over.

I’d echo that approach- a tether drop weight is very similar to the tether management systems giant work class ROVs use - a sort of garage the unit drives out of that hangs from the host vessel. I’ve seen footage of a brov2 used this way with a tow sled, to travel along the bottom at 1000’ !

Having a power supply via tether certainly helps, but going the opposite direction to slim, 2 wire tether can also help with drag dramatically.

With moderate currents, less that 0.25 m/s, any approach is pretty functional in my experience.