My ROV

[a_shorething]

Dude that's awesome (and hilarious).

I love your design idea on that.

I haven't personally tried these, but I saw some video glasses listed with my favorite geek site Adafruit: http://www.adafruit.com/products/1452

They test all of the stuff they sell and critique it. They specifically said the quality of display was good but they were uncomfortable as-is. (They suggested maybe mounting them in ski goggles for wearing long-term, but a welding helmet is bloody perfect for on the water).

Good luck and I'm looking forward to future progress reports on this one...

[kenl]

a_shorething, those are pretty much the same glasses, I have since read a buyers guide on ebay which has provided more insight to the whole topic, they say that the real low end ones are pretty much junk, I'll just put this one down to experience (experience is what you get when you don't get what you expected to get)

Anyways on the day I got some more video, most of it boring, I've cut it back a long way to make it kind of OK but I'll forgive you for opting out early.



I have hooked up my hobby king E-OSD so now the voltage of my 2 battery packs are on screen, it's completely plug and play so no brains on my part.

I'm working towards a real OSD that gives compass heading and a horizon line, with my current testing I don't think I could drive my ROV without spending as much equal time actually looking at it, as I do the screen. To this end I am waiting on a GY80 http://www.ebay.com.au/itm/111015814215 ... 1439.l2649
and also a minimOSD http://www.ebay.com.au/itm/161099711509 ... 1439.l2649

In the mean time I am trying to outsource the coding because it is way beyond me, the best reference I've seen is by bigbadbob on this site viewtopic.php?f=16&t=525&p=6349&hilit=osd#p6349 Once I can work out what I need to know I will PM him, as he seems to have gone quite on here of late.

[rossrov]

Opt out early?? That footage could sell a commercial ROV!! Especially with the subtitles!

[kenl]

That's funny Ross, I'm definitely not looking at going into business, but I do love a challenge and boy oh boy have I got one here.

So many ideas and how to implement them, this C++ hasn't clicked yet but I'm still trying. The Nirvana is coding a complete position hold by scourcing data from the pressure sensor and the 10DOF IMU, although I don't know if the sensors on the chip are sensitive enough?

So, I had a test dive in open water on Monday, first dive to 6m (20') and second in 15m. First up, no leaks, thank goodness. I skipped a step and didn't test the hull to depth without electronics first, so I could have been changing my forum name to 1floodedROV

As I keep harping on about, I had no idea where the ROV was in relation to the boat, and when it started tugging on the tether it was difficult to guess which way to turn so as to stay within the working radius.

The sun was some help, I used to do alot of diving with a Hooka (air supplied via a hose from the surface) and of course the best ground was always just beyond reach. (I had a 120m long hose)

I ended up with a corrupted video file so this video is a bit abrupt as I have only recovered some of it so far, but at least the visibility was better.



The reason for the corrupt video is due to a sudden power outage, I think? I lost all video, but to be honest I'm not entirely sure if I had lost control? Thinking the worst 1floodedROV I hauled it up immediately.

Two of the props were heavily tangled with weed so I think I may have tripped the 50amp circuit breaker I have in there, it's an auto reset device so I couldn't check it to confirm my suspicion.

[rossrov]

The "where is it?" issue is what got me started on my 2 latest ROV pursuits, inertial navigation/improved control and Sonar-related technology.

Like you I am unsure about just how well or otherwise a multicopter-type hobby IMU would work in the slow-motion underwater world. A couple of metres drift in the air is not noticeable for most things but underwater with much lower visibility puts you a world away. Also, above water, GPS can fix alot of the drift problem.

If I actually start building my first practical ROV, then the logic will be to start off with good hardware (hull, tether, thrusters, operator controls, some sort of nav. assistance) and over time work up the code. Something missing from the last one was a programming connector that can be accessed without compromising the hull integrity

[Sunktwice]

I'd imagine use of an accelerometer, and a tilt compensated magnetic compass would give a good response in a position hold mode. The coding, I'd imagine would be fairly simple from an I2c point of view - it's just coupling the two together to create a safety fence and then feeding the calculated drift to calculate the correction value.

Basic data feed from the compass would be a theoretical directional/position indicator, and the accelerometer would give an adequate feedback of relative positional drift - in theory......

[a_shorething]

I'd imagine use of an accelerometer, and a tilt compensated magnetic compass would give a good response in a position hold mode. The coding, I'd imagine would be fairly simple from an I2c point of view - it's just coupling the two together to create a safety fence and then feeding the calculated drift to calculate the correction value.

Basic data feed from the compass would be a theoretical directional/position indicator, and the accelerometer would give an adequate feedback of relative positional drift - in theory......

I would think an accelerometer would only work in relatively still water with no current though, right?

[Sunktwice]

True. You'd need to couple the Accl and tilt compass with maybe a depth sensor aswell. If your planning a hold in strong currents, maybe code in some additional "slop" in the accelerometer deadzone to compensate for more severe pitch and yaw, and perhaps tighten up the compass movement threshold ?

This is all just my inane ramblings - although now i'm curious as to how to get a stable hold myself - I'm not planning huge depths, or strong offshore conditions, so i'll never use it - but i'm very curious now.

[a_shorething]

Here's a thought: Check out the XBOX Kinect stuff. They have some robotic control code that can react to images like tracking a line. If you programmed it to hold position and could see something like an artifact or just some detail on the bottom, it could adjust with throttle adjustments until the image matched again (if it's smaller move in, if it moves left apply right thruster....).

Only works once you can see something, but why else would you want to hold position?

[rovdude]

Awesome ROV!!! I'm working on a design for a second ROV that can achieve deeper depths (not sure what yet), and structure wise it is going to be similiar to what you're doing. For your backdoor(by the way, nice machining ), I see you have two or three O-rings on there. How exactly do they work? To me it looks like when the backdoor is screwed in, the O-rings fit snugly between the interior wall and the aluminum. I've never worked with them, but definitely seems like a very good solution to many pressure related problems.