Just curious if anyone here is doing autonomous drones or drones that are controlled via acoustic modems. Just wondering what is going on in the untethered underwater drone area right now.
Cheers,
-Matt
1 Like
I’m working on an autonomous drone when capacity allows (which in practice means on and off at irregular times).
I have purchased six Nanomodem V3 (NM3) acoustic modems that can be used both for communication (1-to-1 and swarm) and positioning (needs 3 modems with known position).
These modems are developed by a team at Newcastle University in UK and sold by Succorfish.com.
The intention is also to use OAK-D depth cameras with embedded AI processing to assist in the autonomy. We might need to use other camera sensors than the standard to get better low light sensitivity.
Depending on the task/conditions we might be able to operate using cameras and acoustic communication/positioning only, but a DVL and multibeam sonar is most likely also needed.
The shape of the drone will be more like an AUV than a ROV, but with enough thrusters to achieve freedom of movement similar to ROVs when needed.
To ensure long operation times, the battery is large compared to the total size/weight. I have built a 1.7kWh battery for initial tests, but it might be extended if we see that it greatly expands possibilities with few downsides.
A high capacity inductive charging solution is also being developed to enable subsurface charging of the drone for long deployments.
1 Like
Cool project. I’m wondering if you have considered any kind of emergency surfacing solution in the event that power is lost to critical navigation components, or water intrusion is detected, or other disasters.
You might be interested in the Enter Surface Mode failsafe action (that can be assigned to trigger on the Leak Failsafe, for example). In addition, if the vehicle is positively buoyant in the water it’s in, the Disarm failsafe action (or just losing power to the vehicle/thrusters) allows the vehicle to float back up to the surface 
What happens if whatever is supposed to run those failsafes is unable to do so?
Depends on the vehicle. As mentioned, a positively buoyant vehicle that loses power will float to the surface. Conversely, a negatively buoyant vehicle would sink.
Leak sensors are generally expected/designed to detect a leak before it starts impacting major systems. As an example, if water starts coming in past an enclosure end cap then a leak sensor placed near the end cap should get wet before the electronics do. Similarly, leak sensors placed at/near the bottom of an enclosure should get wet by rising water before it rises to where the rest of the electronics/connections are.
Depending on the nature of a leak, slow leaks should generally reduce as the vehicle gets closer to the surface, because there’s less pressure forcing it into the enclosure. A highly dynamic ‘leak’ (due to a projectile hitting and breaking the enclosure, or the enclosure imploding under pressure) is much harder to handle gracefully, but unless all your electronics are encased in epoxy or something then that’s generally not a scenario that’s expected to be ‘handled’.
If pressure-induced implosion is an expected operating risk then that’s almost certainly intentionally operating beyond the pressure rating of an enclosure, in which case it’s better to use an alternative enclosure design (e.g. thicker walls, metal instead of plastic, etc) than it is to try to introduce safety mechanisms that can automatically recover/surface an imploded and fried vessel.
If you’re concerned about the electronics themselves failing unexpectedly then the effects of that depend significantly on the electronics in use, how they’ve failed, and the active control system (assuming that’s still functional). I won’t do a full AUV operational risk analysis here, but if you have particular scenarios you’re concerned about and want to know how ArduSub or some other system would or should respond then I’m happy to hypothesize/discuss
Thanks, Eliot. Lots of good points to think about.
1 Like
We are using multiple sensors and using sensor fusion to create a multispectra representation of the system’s surroundings. We have the good fortune to be operating in very clear water where lights can provide significant viewable area. The project is moving slowly in part due to me but we have the vehicle all working some sensors integrated but the major sensor fusion and then the navigation is very behind. Ohh… I should mention we are using a number of neural networks and some simple machine learning to help with navigation and of course sensing and sensor fusion. Last night I also finished the software and hardware for our mesh-enabled local communication network that allows multiple devices to communicate and more importantly divers to work with AUV. Mesh coms system is neat. Each node uses RF, we plant to allow acoustic later, to communicate and we get decent range but since each node is like 150 to 200 dollars in parts, we are looking at using different components to get it below 100, you can flood a local area with them, get high-speed comms as well as good localization without GPS. The mesh also continues to work fine each when nodes are lost or lose connectivity. I’ve been using them to stream sensor data, and video, in difficult areas. A partner of mine also designed a dropper for them but we will most likely have to design a new drone to deal with the additional hardware the idea is the system can make bread crumbs or provide better comms at the request or at need. Sounds like we are doing similar stuff. Perhaps we should compare notes sometime.
1 Like
16 posts were split to a new topic: Sucorfish Nanomodems
@Robi_Sen
Thanks for the information regarding your project. I would love to hear more about it. This is maybe something you are able to share openly here in the forum (so that more can learn/be inspired from your project activities)?
We haven’t had much time the last months to work on our project, but I hope to get it prioritized next month.
Hence we don’t have a lot of actual implementation yet, but we have researched a lot and tested many of the individual components/parts of the planned solution separately. I have no issues with openly sharing what we have got, but I’m really bad at documenting the progress I have made. Luckily my colleague is much better at documentation and is pushing me to improve
Just curious about your streaming video quality. What resolution and FPS are you able to transmit underwater?