So I’m brainstorming ideas for an underwater drone I’ve been wanting to build so I can mess around in the great lakes.
In the long term I’m looking to send a reasonable resolution video signal (or two) over a tether as well as sending the necessary data to control the drone.
I’m trying to figure out how I’m going to get a video signal back to the surface in real time so I can see what’s going on down there. I’m planning to have the drone record the video locally to a hdd so I can later retrieve the high quality footage, while at the same time sending a probably lower quality signal back to the surface for navigation.
For the drone tether I was planning on just using an ethernet cable, the maximum depth in the great lakes is in lake superior at about 400m so 450m is going to be the long term goal. Ethernet really only goes to 100m but it looks like I can get a few repeaters to get the necessary length.
I’m not really sure how I should approach this. I’d like to keep the computer hardware on the actual drone to a minimum. Originally I figured a rasberry pi would be able to handle it but if I’m going to be transcoding video signals so I can send them up the tether I imagine I’m probably gonna need a 1030 or something just to handle the transcode. How do you even send a video signal over ethernet? I’m sure it’s possible since I’m pretty sure security cameras don’t require you to run an hdmi cable all over your house they just need access to the network. Anyone have any idea how much of a problem signal degradation is going to be?
If the drone camera’s have an IP output and you have room (or you can stream video over IP) , you could get a media converter that would be powered off the drone’s power system and then run multimode fiber up to 500M or single mode fiber for kilometers. Here’s an example of an industrial rated media converter. 10-100-1000 Industrial Converter | S-1110-XT Media and Rate Converters
Also, not a recommendation, just a google search.
You would want to protect the tether in some way like:
On encoding, the Raspberry Pi does have a GPU capable of realtime 1080p h264 encoding, and if you are using RPi’s camera, then the camera itself can produce an h264 or mjpeg stream.
This can then be sent over the fibre cable using one of many network streaming pieces of software.
a 400M ethernet cable would probably weigh about 40lbs all told. But it’s hanging out in water so it shouldn’t be too much of an issue. I’ve built a drone that has a tether that’s probably 15lbs and 50ft or so. Basically 8,12 gauge wires that are tied together into a cable and the drone wasn’t bothered by the tether at all. As long as the tether is relatively close to neutral buoyancy and I’m not stretching the cable to it’s limit I don’t think I’ll have an issue.
This is fascinating, I wonder what you find down there (not hoping for Nessie, but I’d be interested).
Are you publishing any of your work in any form?
Ethernet is nice because of POE- you can send power down and maybe even run the drone off of it… or at least just trickle charge the batteries.
There’s these extenders that you can daisy chain.
The extender+case costs about $25-$30.
Mikrotik has some RJ45 stuff that uses slightly more efficient/better/higher power signaling… which is why they claim “… distance of CAT6 Ethernet cable between two GPeR devices is up to 210 m …”.
Regarding weight and cabling I just googled Ethernet cable bouyoncy, and I keep getting shopping links for a this bright yellow neutral buoyancy armored cable… I guess sending ethernet underwater is a thing
I’m also imagining how I’d ghetto DIY a dumb ROV using this setup.
Cable+extenders for data+power
get an ultra low light 12MP security camera
get an usw-flex or some other cheap poe splitting switch
Ethernet connected pi for “avionics” or whatever that’s called
casing + motors + auto homing… hmm, do you build a “float up on power loss” mechanism, or do you build some smart “trace back” mechanism… do you do expansion chambers for buoyancy control like submarines, … or do you do neutral buoyancy and propulsive steering only (e.g. 3 propelers on the back)
how do you navigate, there’s no GPS there… or is there? sonar?
As far as publishing results goes, basically that involves throwing up the recorded video onto youtube. you can search TomSmash on youtube and ive got a couple videos up of me testing my drone. If i start grabbing temperature data and pressure or whatever else i end up getting ill probably publish it if its not too much of a hassle. Itd be pretty cool to find some old shipwrecks, theres plenty in the great lakes.
Right now i have a very basic drone that has no electronics on it, except for a go pro. Basically a pvc frame, 3 sump pumps that i attached propellers for thrust, some lights, a control panel consisting of 7 toggle switches and a car battery to power it all. 3 Toggle switches toggles motors on and off, 3 toggle switches control forward and reverse, and a toggle for lights.
Float up on powerloss seems a little unnecessary. As long as the tether is strong enough and the drone small enough, just reeling in the tether should be enough. Likewise navigation isnt really an issue at least how i anticipate using it. Im imagining me mostly using this from a boat and sending it straight down to the lakebed.
Definately trying to manage feature creep at this stage. Next goal is basically making a drone that can sit at a depth of 100ft for like 6 hours and not have any leaks. Also for this version id like to be able to pass a video signal to the surface so i can see whats up down there. I’d like to focus on accomplishing those goals befpre i start adding ballasts for buoyency control or more sofisticated motor setups. At some point id like to be able to control it with an xbox controller or a hotas instead of manually turning motors on and off. Plenty of work ahead of me thats for sure.