We finally got all of our lab’s BlueRobotics vehicles running simultaneously! Off to deeper waters to test our USBL system next.
(Yes I was terrified of chopping a tether the whole time)
Hi Sam, can you share a bit more about your configuration and use cases? We are building a coupled system with a BlueBoat and a custom made ROV for algal bloom monitoring, and a winch to manage the tether, so interested in finding out more about yours.
Hello! I’m using these vehicles as a test platform for distributed control algorithms. I research how we can use graph neural networks (neural networks distributed over a network of agents) to control vehicles in environments with limited communication. I’m working on getting these ROVs + Boat to be a physical platform to test these controllers. I’m also working on getting acoustic modems integrated into the physical vehicles to make our limited communication setup more “real”. Alongside getting these real vehicles working, I’m developing a HITL simulation cluster (“Raspberry Pi supercomputer”) that runs an additional 10-20 simulated vehicles that run alongside the real ones.
In terms of physical hardware on these vehicles, the ROVs have a Ping Echosounder and a Cerulean ROVLe MK II USBL receiver. The boat has a USBL transmitter mounted on a boom that we can lock down.
As for the configuration of our hardware as it is now, everything is run though a central switch. I modified the FXTI boxes and the BaseStation to accept plain old Ethernet cords. Those get plugged into a switch along with one laptop per vehicle (and another for debugging and monitoring). We manage the tethers using Reel-a-pails that we modified to include a slip ring.
Tether management during experiments is definitely a concern for us and we’re still playing around with different solutions. I don’t think letting the tethers float on the surface with the boat in play is a great idea. We’re going to be prototyping some tether-excluding propeller sheaths / ducts, as well as testing a tether weighting system that biases the tether towards the floor.
Hi @SamH -
Give driving over a tether with a BlueBoat a shot - I think you’ll find the keel protects from entanglement pretty well! In my limited testing of a towfish it hasn’t been an issue thus far…
Really cool to hear about your setup, thanks for sharing those details! I’d not heard of “Reel-a-pails” before, handy!
What acoustic modem do you plan to use? Something DIY or off the shelf?
I’ll give it a shot. I’m not too worried about driving perpendicularly over the tether given how far the skeg extends past the propeller, it’s the parallel-ish approaches that give me pause.
For the acoustic modems, I think it’s going to be custom/prototype hardware. Next school year, I’m helping advise an undergraduate capstone project whose goal is to integrate a modem into the BlueOS ecosystem and establish MAVLink communication with vehicles and a ground station over acoustic comms. One of the students on the project has industry connections to an acoustics company and is building some sort of prototype this summer as part of an internship.
Hi Sam,
Thanks for the overview of your awesome project, sounds like an ambitious undertaking. Ours is more focussed on single ROV and USV connected by a tether so that we have a means of real time data transfer to cloud from the environmental sensors attached to the ROV. We are focusing on building a rock-solid winch system that can keep the appropriate amount of tension on the tether such that there is enough to allow free ROV movement with minimal slack tether floating around on the surface. We have a Waterlinked DVL so will use this to determine relative position of USV and ROV. We have some 3D printed mesh guards on the BlueBoat props as well as the ROV thrusters to minimise risk of tether damage. The project is in the early testing phase so will probably post on the forums once we have a bit more to show.
If you want to collaborate further, let me know. We are part of Adelaide University (https://adelaide.edu.au/).
Regards
Eugene