Hi Kevin,
Sounds good. I’ve been doing some test prints with the 6" hull files scaled down to get some print settings dialed in. I’m using PETG and it’s working great so far.
hi，guy.It seems a good choice to add a rudder to the stern with an underwater servo to control the direction.
Hi @octopus , do you know underwater servos that are cheap and reliable?
@BayesianMonk Blue Trail Engineering makes very nice servos.
1 Like
Hi @kklemens We are Team Amogh a group of students from Indian Institute of Technology Madras working towards developing an torpedo shaped AUV. I came across this thread while researching about AUVS. If you are open to discussions I wanted to discuss some stuff regarding the hull. Could you give your personal email address or any other contact ?
Looking forward to hearing from you
Thankyou
Hi @TeamAmogh thanks for your post! Sent you a private message with my email.
Hi all,
regarding the firmware to control a torpedo like AUV, is there a ArduSub version that can do so? I am reading threads from 2018 but it still not clear.
Hi @FraSec, welcome to the forum 
The design Kevin presented here is specifically using only fixed-orientation thrusters, in which case a build of ArduSub with a custom frame should be sufficient for controlling it, provided it has sensors that allow it to estimate its position and the like.
If you’re after a more traditional AUV design which makes use of control surfaces then that is not currently possible with ArduSub, and developing a compatible version may be more challenging than adapting a firmware like ArduPlane to work underwater.
These guys have an interesting build series on making an AUV with a bunch of Blue Robotics parts. They say they use a Pixhawk, but not which firmware or frame configuration.
Hi @kevink !
It is indeed a cool project! We’e been in touch with the CPS drone folks, you may have noticed the affiliate link in the first video. We were hoping to have them test some of our upcoming servos, but it seems they don’t need them anymore…
They are using a pixhawk, and some python scripts, and potentially ArduSub if not ArduRover… fun stuff!
Thanks for this Kevin
Nice to see someone going on and developing a “consumer” level AUV, and great to see them ditching the rear fins.
I did over 100km of scaning with a Seaber unit about 12 months ago in tropical reefs systems and it flet like every 3 mission failed due to obstical avoidance as the fins (limited to a max acent rate of 15 degrees) couldn’t react fast enough when we ran near sand reef interfaces. I hope they get the batteries out of the electronics tube to allow for easy swap out (and more importantly allowing the batteries to be broken down so they can travel on aircraft)
1 Like
Hi @Scott_W goot to hear from you! Things have certainly advanced since we started.
I looked into a Seaber too last year and got a few quotes. Its pretty expensive to purchase with a side scan and nobody could tell me how the insurance worked on the rental side (which was expensive too). I reasoned that I might as well keep researching on how to build my own, so if I lose it, that’s on me.
I’ve actually been out operating Remus 100s lately for work so I have an idea of what is important and what is not.
The battery situation is super interesting because you need a lot of cells and they need to sit low in the hull. Bluefin Robotics does batteries in modules that are easier to remove, but I heard they have leak issues. Even the big boys have have issues.
I have become partial to 21700s as they are fairly easy to source and pack a bunch of power in a slightly larger size than an 18650.
ArduPilot has support for third party INS systems now so I am thinking about how to scope the requirements for my vehicle.
I need to scan at 100m depth so I definitely need to be able to start on the surface with GPS, hand off to the INS when in the water column, then transition to DVL once it gets a lock.
Hi Kevin
I suspect many if not all of the thing on your “what is important” list are pretty much the same as what’s on mine.
RE the DVL navigation I found we could only trust it for around 1-2hrs (Seaber were using a WaterLinked DVL A50 as after that distance the survey would be too far off (1% of distance travelled). Although the unit we had was fitted with Deep Vision Sidescan where we were able to swap over the transducers from 200kHz (best range in real life we got was 70m [140m swath]) to 600kHz (best range in real life we got was 35-40m [70m swath]). So we did longer runs to scan the areas with the LF transducers to scan the area, than swapped over the transducers to the HF transducers and did shorter runs over the targets to pinpoint and gain detail, which worked quite well
An aside on Batteries there are almost 2 different issues here
-
Travel on aircraft – The 100Wh limitation – I get around this with the ROV by having a couple of smaller batteries at this lesser capacity and hot swapping them on the boat. But for a AUV with more capacity this become a huge issue air freighting Dangerous Goods (for me typically multiple flights to small pacific islands) and is near imposable and costs a fortune.
-
Not having to open the battery compartment for charging (ie minimisation of risk of flooding)– It still sort of surprises me Bluerobotics haven’t implemented a charging equalization balance board in the battery compartment and a +ve and -ve charging port to eliminate having to open the battery compartment (sort of common on a lot of dive torches now)
Scott
That’s super interesting on the DVL, but about what I would expect with the 1% range error over time. I think for the technology level, you just have to plan for a GPS fix about every hour, which is reasonable.
I would probably go with a Cerulean Tracker 650 for my build, to keep costs low. I could deal with the size.
The swappable transducers are an interesting concept, but I would have expected better range on the 200kHz.
I would probably go with 2x of the Cerulean Omniscan 450SS transducers. I have had good luck with one on the front of an ROV, even if the operation is a bit weird to use it. Shouldn’t be an issue though for an AUV.
One piece that is missing from all this is an acoustic communications method. Aborting a mission or just finding out where the vehicle is, is critical. I don’t know of any hardware yet that can do MAVLink. There is the ahoi Project that may work, but definitely going to take some software work.
Hi Kevin a picture tells it all
This is the same object an 3 different passes, the object is around 15m long but less than 0.5m above the seabed
The swappable transducers were really good (just a standard subconn plug and a recalibration run). It saved a hell of lot of diving targets (over 150)
The Open-Source Acoustic Underwater Modem Ahoi project looks really interesting, but given the choice of where to spend $'s I would go for a longer range short baseline acoustic positioning system to better correct for the drift - maybe something with around 2-3 km range (prompt waterlinked) rather than comms to the unit
Super awesome project! Trying to learn from open projects out there. I have a similar project in mind. I’m planning to make everything public, designs, CFD simulations, coding of the kalman filter, etc. It’s partly a learning experience for me. Plannig to use 150mm housings from BR (if they become available), T200 initially at least, etc.
Last quote I got from L3H for an Iver was > $200k USD for mostly a basic unit.. and the main issues is that if you add any sensor the price becomes incomprehensible (DVL, acoustic modem, INS, SONAR). Hence I’m thinking of a simple design using WaterLinked DVL as the most expensive navigations sensor. At a minimum have a DVL, INS (MEMS based?), and pressure sesor. Do you guys have experience navigating only with these package? How bad/good is it? – the alternative is to add a higher grade FOG based INS, but this adds a lot more.
Is the a maximum limit of battery you can carry? I’m assuming it needs to be broken to <100Wh each pack?
Hi @bryan3D
Have a bit of a look through what I am suggesting over at this post specifically with the US$200K in mind
Specifically in relation to your question in the post towards the bottom I show a real world 7.63km mission with a Yuco (with a waterlinked DVL) which when it surfaced was 98.9m away from where it though it was (so 1.29% error over the 59 minute mission) there was a little bit of current there (nothing too bad) and was about the positional error we were consistently getting over the 18 days
I’ve mainly played with the Yuco (a short hit with the Remus) I think @kevink has been playing with ?Remus’s ? (he told me but I can’t 100% remember the brand he was working with) I think it’s Ok and you can work with the cheaper limitations of just DVL (no LBL SBL style system) although for anything significant that you are chasing data for I would suggest doing a rescan of the target at the beginning of a different mission to tighen the location
Scott
I have travelled with I think 6 sub 100Wh batteries and had a chat with a supervisor at check in (I always tend to be the anoying one that blocks the check in queue)
Blockquote From The International Air Transport Association (IATA) is a trade association for the world’s airlines. It was founded in 1945 and currently represents around 300 airlines
Note 2 Each person is limited to a maximum of 20 spare batteries of any type. The operator may approve the carriage of than 20 batteries
I always have the batteries out and present them taped up and have copies of the batteries paperwork and copies of the guidelines to give to the person at the check in
Scott
1 Like