OK, hello all, let's kick it off. Welcome to the Hack Chat, I'm Dan and I'll be modding along with Dusan (I think) as we talk about underwater robotics with Tony White. Not sure I saw him log in yet, though -- are you out there Tony?
Hi everyone! Happy to join you today and talk about some cool stuff! A bit about me -
Generally I consider myself a jack of all trades, with a speciality in rapid prototyping subsea and ocean surface applications. I’ve always enjoyed the water, followed that thread throughout my life, and have really grown into engineering in the medium with scuba diving and all manner of waterborne electromechanical systems testing. From recovering sensors to mapping surf breaks, so much is possible!
After doing autonomous survey vessels in Louisiana / England with C&C Technologies / ASV Ltd., I went off to LA to develop swarm robotics and began working with Blue Robotics! This is a fun video too: https://www.youtube.com/watch?v=zNYzljKr_wY
Little bit of aside but still relevant, might kick off some discussion: We were talking in our editorial meeting yesterday about the submersible Alvin, which was built in 1964 and is still in operation? How do you think is that even possible?
It was built tough - titanium spheres, and lots of expensive materials. That said, I'm sure much has been renovated on it and refreshed over the years, particularly the controlling tech. It's capable of reaching some pretty incredible depths!
That's what we figured -- probably everything on the original vessel has bee replaced over the decades. Still, it's pretty amazing that it got this far
for sure! The ocean is pretty unforgiving, especially when it comes to corrosion. I'm more impressed by ships / boats - they exist at the most consistently violent place on earth - the sea/air boundary, which never stops moving! And many are in service today that are older still!
I'm not an expert in the materials science around the titanium used, but I do know that much of the titanium used in AUV spheres comes (came?) from Russia, which has been tough in recent years for the industry.
Hi Tony! Blue Robotics has been doing some amazing stuff with regards to lowering the barrier to entry for the underwater community. Which advances/improvements do you see coming into play in the next ~5 years. Alternatively, what should we be focusing on to improve/lower these barriers more?
Hi Fred! Great question. We've got a lot of exciting products in the pipeline, I think our wetlink connector series is the biggest game changer. Letting users terminate their own cables with connectors, with 0 potting involved, for <$100 will shake things up vs. current subconn connector practices < td>
</$100>
I've worked with ROVs, which use a tether, and autonomous surface vessels like the swarm I linked above, and of course the BlueRobotics BlueBoat. I've been developing a hovering AUV recently, hoping to get it in the water soon! All based on BlueOS / ArduSub
As for "cyber resiliency " - it's uncommon to have an internet connection when in the field deploying an ROV! Nothing like an air-gap to keep a digital system safe....
What about the communication protocol, Tony? You guys thinking about how all those ROV devices (thrusters, lights, grippers) will communicate in the future.
Thomas I'm not sure if titanium is in the nodules that everyone is talking about harvesting from super deep in the pacific - mostly cobalt and other battery metals I think? Quite a tough environmental question
GitTyler805 - we typically use TCP/IP over a twisted pair via homepluf adapter - like the powerline ethernet adapters used in many homes. We're working on some exciting stuff around 10baseT for similar communications between vehicle payloads (lights, grippers, etc.) and hope to upgrade our tether electronics to support 500m + at gigabit speeds. Currently our solutions top out at 300m / ~80mbps
I saw an approach used by AUVs, where the battery pack is on a linear stage, which i believe helps it dive at certain angles, is that something you've used? (hope this message isn't repeated, got a failed to send error)
Our ArduSub docs do indeed need an update - it's coming very soon! We have been updating the docs here quite a bit: https://blueos.cloud/docs/latest/usage/overview/
@Clamb - tether management is often a huge consideration! If going super deep, a "clump weight" attached to the tether some distance from the ROV lets the ROV not have to fight tether drag. However, you can't go very far from the weight as a result! Our slim tether is a great way to reduce drag, as it only has a single twisted pair but is just as strong thanks to the equivalent amount of kevlar
Hi Tony!
Thanks for the insights into the air-gap as a way of mitigating cyber threats. I'm interested in pulling a bit further on the cyber resilience thread. I assume that these underwater systems are equipped with a range of sensors and communication systems for navigation, data collection, and decision-making. Given that these systems are critical for their operations, I imagine they could also be potential attack vectors in certain scenarios.
Could you share more about how you address these vulnerabilities, particularly in terms of resilience against threats like GPS jamming or spoofing, and any interference with acoustic communications? How does the system handle situations where these communication or sensor systems are compromised? Are there redundancies or fallback mechanisms that ensure the robot can continue its mission reliably in such cases?
I’d love to hear more about the strategies you employ to ensure these systems are robust, even in the face of cyber-physical disruptions.
@darkmoon3d no, we moved away from potting years ago! Our WetLink Penetrator is a mechanical gland that can seal the correct sizes of cable to 1000m! https://bluerobotics.com/store/cables-connectors/penetrators/wlp-vp/
Re barrier to entry... What about getting rid of pressure hull need? Fill the insides with gelled oil (candle gel is a fairly pure paraffin oil gelled with some styrene copolymer, a friend tested it for easily remeltable high voltage potting). Can be easily broken down for maintenance/repairs and then remelted again. Heard submarines use the potting-in-oil trick for stuff outside of the pressure hull, and the gelling trick may make it less likely to do accidental spills on the carpet. For the propulsion, what about motors with noncorroding ceramic bearings and fully potted coils, eliminating the need for moving seals?
@anfractuosity I've seen moving batteries primarily on AUVs, buoyancy gliders in particular. Typically the Blue ROV2 doesn't angle down to descend, but just thrusts itself downwards like an upside-down quadcopter. I've not messes with a linear stage like that before directly!
Hi @Thomas Shaddack - many deep deep vehicles do use oil-filled enclosures - the advantage is depth rating, but working with them is always a pain! Also, the oil in the enclosures will reach the external pressure thanks to a bladder that is compressed, keeping a 0 differential across the seal. This does mean that all your electronics have to be pressure proof - the small air spaces in mems devices like accelerometers are crushed at extreme pressure as a result! Therefore air-filled rigid enclosures are the best bet in most scenarios. We are starting a project with WHOI at Blue Robotics that may result in more affordable 6000m rated enclosures coming to the market someday!
sending this again as I got a didn't send notification: Hi @CPSfarmbot - GPS isn't an issue for underwater systems - radio waves don't penetrate water! Typically acoustic localization systems are used, that may tie to a surface GPS. Generally, if the right acoustic frequencies were disrupted the vehicle would be helpless. It sounds like your question is more directed at AUVs, independent vehicles, that don't have a human a the controls?
@Thomas Shaddack as for motors - that's actually how the Blue Robotics T200/T500/M200 work! We use oil-impregnated plastic bearings rather than ceramic, and the stator is potted, so no seal is required. When connectorized, we've had users use them at full ocean depth !
@Tony White , my understanding was this chat was about " about the harsh engineering realities of underwater automation" and not just BlueRobotic tethered systems. From this perspective there are numerous vectors that should be consider for system resilience. I would even go as far as saying that even tethered systems are susceptible to cyber-physical compromises from the Repo to the Sensor. Some food for thought. Thank you for your time.
there are many different types of potting compounds. Some are more flexible then other after cured, which can help with adhesion. Some can cure underwater, like loctite marine epoxy. Some are more thermally conductive! Generally, the environment they are cured in is critical to the performance, and if exposed to warm/cold conditions afterwards failures can result. O-rings on the other hand, always work!
Isn't it easier to put the MEMS things into tiny sealed pellets instead of having to maintain a whole big pressure hull with leaks? Small objects like this could be also easier made redundant to cope with seal failures.
Also, for garage lab tests, the pressure washers could be used for depth testing up to approx. 1 kilometer of simulated depth. They are used in DIY contexts for things including hydroforming.
@CPSfarmbot for sure - not trying to keep the discussion centered on Blue Robotics tethered systems, they are just inherent to the ROV genre! Surface vehicles are definitely vulnerable to RF jamming, the nice thing about acoustic jamming is that the power required means that the radius of effectiveness is considerably more limited? I generally have avoided all military aspects of the field in my career, as I'd rather not contribute to anything that brings people harm
Are there ways to keep a mesh network between independent underwater vehicles? A way to monitor their relative positions? Eg. for autonomous seabed or shipwreck mapping.
@Thomas Shaddack Pressure washers are a great high-pressure source for test tanks! Definitely used stuff based on them before. It may be possible to pot MEMS to give them more strength, but liquids under pressure have a way of separating such bounds at a tribological level. The fastest route to prototypes is usually chucking the electronics you want in an enclosure and going from there!
As for mesh networking - acoustics have more limited channel bandwidth due to the low frequency (again compared to RF) but some of what you describe is definitely possible. Cerulean Sonar makes the ROVL, which supports multi vehicle use. https://docs.ceruleansonar.com/c/rov-locator/multi-unit-operation-swarms
Hi all! I'm involved in a swarm entertainment water robotics project called "hydrones". I was wondering if there are any off the shelf, cheap. ultrasonic coms systems for underwater use?
@amok.mcconnell That's a broad question! Shaft seals are generally easier for rotary than linear applications. By joints, what sort of mechanisms are you thinking of? Avoiding galvanic corrosion is key, but generally water provides nice lubrication for most applications
Could an array of such transducers, on the vehicle surface, act as a de facto 3d imaging sonar array (incl. passive modes)? The ability to focus a beam would be also helpful for communication. (More thought. Could the communication of one vehicle be used for simultaneous imaging by other vehicles, in a way akin to a bistatic/multistatic/passive radar?)
Hi @paul it sounds like you're after an acoustic modem. Something like these may be affordable! https://succorfish.com/products/delphis/ This is also a good reference: https://udrc.eng.ed.ac.uk/sites/udrc.eng.ed.ac.uk/files/attachments/UDRC%2025-03-2021.pdf from this forum thread
https://discuss.bluerobotics.com/t/succorfish-nanomodems/14496
there's an excellent tutorial on instructables for building a cheap diy hydrophones https://www.instructables.com/Lets-Build-Some-World-Class-Hydrophones/ . perhaps this could be integrated into signal processing?
@Thomas Shaddack typically the acoustics used for localization are much different than what would be needed to create a map from sonar returns. The Cerulean Surveyor multibeam (only $5k! ) is a good example of a small mapping device, where as the ROVL already linked is used for localization. Other localization systems use a longer (short) baseline like the WaterLinked UGPS. Remember, acoustic channels are typically Khz, rather than Ghz, so a lot less is possible from a bandwidth perspective !
I was wondering about joints / servos that give a manipulator grabber arm its movement, torque and position feedback. Are the motor bodies housed inside the ROV hull and linked through the hull mechanically or are the motors in the joint themselves?
@amok.mcconnell typically subsea servos are housed in their own container, with penetrations for the mechanical rotation and the cable going to them. Blue Trails engineering makes a great option, and we're soon going to launch one from Blue Robotics! https://www.bluetrailengineering.com/product-page/underwater-servo-ser-20xx
That's why I wanted to combine multiple uses into the same signal environment. Sacrifice some efficiency in comm so the ping can be simultaneously used for ranging/imaging, and you don't have to share the medium with both kinds of signals. Encode data to the pings so the reflected sequences can be used for imaging and the received ones we didn't send ourselves can be used for comm (and if strength allows, also for imaging).
I've avoided using acoustic communication methods in favor of a tethered solution due to cost and simplicity on my own projects. hasnt it also been proven that acoustic methods are harmful to aquatic life?
I was impressed with the design of the swarm-ready "data diver." As much as I keep thinking about gliders and energy-saving, I really liked the way they "power down to depth" by reversing their thruster, and float to the surface when they stop thrusting.
Can you tell me what the teardrop-shaped capsule is all about? (sorry, I know we're trying to stay general and not about your company)
acoustics are typically only harmful to marine life if at very high power levels, in close proximity. This occurs when sub-bottom trandsucers are probing the earth from large arrays, looking for oil deep in the earth. Can pop a whale eardrum easily! The amount of power in the water from a small sonar is typically not enough to cause issues, especially at the frequencies used. Range is also limited as water is 10x denser than air, so propagation losses are high!
@Christopher The data-diver has a clear payload/electronics tube, right next to it is a yellow "organic shaped" outrigger--that's what I'm calling "teardrop shaped"--and at the bottom is a caged thruster.
Hi @Tim McNerney - The data divers were a blast to design! They were actually made with Apium, not Blue Robotics. They can only descend vertically, running one motor slightly faster than the other to "corkscrew" down. We did some cool synoptic measurements of internal waves with USCD researchers off San Diego with them!
The teardrop thing is overmolded buoyancy foam, positioned there for stability.
They were never commercialized, but this company is working on their successor! https://www.jaia.tech/
Thought I had re hydrophones. Poor man's SOSUS. Digitize the signals, add precision timestamps eg. from GPS, transfer on a non-realtime medium like internet (or record and pick up later if we don't need near-realtime processing).
Yup, logging hydrophones and sharing the data later, or in realtime, is possible! I've been testing some hydrophone options for our Reef with the BlueBoat and doing just that - trying to record whale song! I got a bunch of reef sounds yesterday but no whales came by
sorry I give up - first time here and last time - the text jumps too much - can't go back and read what I missed - tryed up arrow and just keep jumping to the bottom - no stinking way to read it - sorry my laptop screen only lets me see maybe three inches by the time the header and "you have" MESSAGE blocks the bottom - maybe this will be PDF and I cam read it next week - but over an hour and I an frustrated and fed up
Using acoustics for both measurement and communication would be possible, but things like a multibeam imaging or echosounder are actually beam forming with multiple transducers, so doing a communications chirp at the same time may not be possible ? The folks at Cerulean Sonar are experts in the field, and always willing to consider new approaches!
Thanks. Having mentored university students, I also think about ways to make untethered failures less heartbreaking. Testing long-range autonomy is hard.
sorry @mikeb - you can scroll the chat window with your mouse, but it does seem to jump down with each new message. I'm not sure about a pdf, but the text will just exist here to read at your leisure!
The beamforming is generally done by transmitting the same message with different delays. (Steering of a microphone array into a direction uses the same trick - I think recording several microphones at once could then allow steering in postproduction, too.) The old Wullenweber antennas even used different lengths of coax as delay lines, I think. The F35 radar is said to be also able of acting as directional radio so it should work in principle.
I'm not aware of any revolutions in acoustic modems - I've been researching some options for the Reef (linking since I keep mentioning https://bluerobotics.com/the-reef/ ) - but I've not heard of near-shore noise issues. Are you referring to reef-crackle, or engine noise from boats/ships?
if you go to a healthy coral reef, and go for a swim, you'll hear quite a crackling! I can probably link a recording from yesterday shortly to illustrate
I wasn't aware this could interfere with acoustic modems, but it doesn't surprise me! I'll have to checkout the FFT to see what frequencies it's affecting
WIFI uses multiple antennas tricks with using delayed/reflected signals as separate channels to use wider bandwidth. SDR radars use tricks with autocorrelation of modulated pulses to gain sensitivity. Could software radio have tricks for signal processing of acoustics?
Re: "reef crackle" or not... I guess I'm referring more to high white noise levels near the shore, e.g. from waves crashing, etc. In the open water, it is quieter.
I think the wetlink penetrator glands are a big advancement to help students @Tim McNerney - they are such an affordable way to bring a cable into an enclosure, and learn alot about sealing in the process! BlueOS in general is positioned as an open-source software pickup-truck to make development easy - with extensions like Node-Red students can be doing simple payload integrations quite quickly! https://bluerobotics.com/learn/blueos-and-node-red-guide/
And Cockpit - an open source ground control station software makes customizing your drone control interface for any type quite easy too! https://github.com/bluerobotics/cockpit
here is an example of that "reef crackle" recorded yesterday with an ACsense hydrophone! https://drive.google.com/file/d/14VRU9RANEXXzHPKZebg8GuOd_YLg2t7b/view?usp=sharing
https://acbotics.com/acsense-mini
Re the acoustic imaging, thought about a thing for divers for zero-visibility. megahertz transducers around the goggles, imaging within few feet, showing augmented reality image. The same could be, if it would work, used for robots to feel the space around them.
@OceanLab Wow, these WetLink Penetrators are quite inexpensive. We were using waterproof connectors but never tried to solve the seal against the enclosure problem, as we were mostly staying near the surface.
Very cool cooncepts @Thomas Shaddack ! A multibeam imaging sonar provides pictures for such situations, and I've seen some examples of products targeting divers with them. They typically cost many 10s of thousands of $, and so are limited to military / oil and gas applications typically? The Sonoptix Echo is $10k, and is the most affordable multibeam imaging sonar out there! This guide does a great job explaining the differences on sonar types: https://bluerobotics.com/learn/a-smooth-operators-guide-to-underwater-sonars-and-acoustic-devices/
Connectors are always going to be more expensive than a gland, as there is a lot more that goes into them. But for situations where you don't need to break the connection, they are great!
Do you have any advice on how to make reliable variable-buoyancy devices? (as a diver, I keep wanting to say "BC"). There are many ways to build these, but keeping it simple is super important underwater.
The silicon dessicant sometimes does just this! I'll use a papertowel if I know I have a leak and need to mitigate it for a short time - preventing sloshing is absolutely key
There was a study of shark skins by James cook university 20 years ago, on the use of surface microgrooves to deter barnacles on ships. It kinda works.
@shaddack The marine industry has been dealing with bio-fouling forever. As Tony points out, it is the toxicity of anti-fouling paints that has become an issue more recently. I've never tried building a robot entirely out of copper. Is that disallowed too?
Thought. Controlled release of metal ions from surface by electrochemically modulated corrosion. Something like sacrificial anodes but the other way, and only when needed and where needed.
The manufacturing process is basically an embossing of plastic using patterned micro-etched plates. It would need development to be adapted for odd curved surfaces.
For anyone still here, I just started this livestream. I'll try to monitor both this and that chat at the same time! https://youtube.com/live/3WR2bp3m36g
Thanks for all the interesting chats folks! Feel free to reach out anytime - tony@bluerobotics or via our support form https://bluerobotics.com/contact/#contact-us-form
Random thought re the tether. Ukraine battlefield uses drones with optical fibers to work around jammers. Can the same tech be leveraged, possibly even as a surplus once the hostilities end and the toys hit the markets?
Lutetium
