i think i want to marry him.
that's a seriously brilliant idea.i'm in for 40% of the profit.
(Fuckpig "That's not even a pound a midget.", Fri 31 Oct 2008, 12:22, 1 reply)

It would be brilliant
if it actually could work. I've given this thought for years now and have concluded that it's not feasible- setting aside for the moment the construction of neoprene harnesses for fish, what sort of sensors would you use? How would you control speed? Could you make the sensors small enough that the fish could see around them?

It would be an engineering challenge that would tax the ingenuity, to say the least.

He's come out with other ideas which are far more practical. If I can make just one of them into something workable that will make money, I'll set it up to go completely to him.
( The Resident Loon Not a demographic. Do not measure., Fri 31 Oct 2008, 12:32, closed)

Engineering challenge?
You'd have a small optical sensor (small LED & LDR pair, something nice and non-contact like that) detect the position of the fish's tail and fins and hook that up to a computer through an ADC. You'd then know the position of the fins, meaning you'd be able to figure out the direction that the water was going to flow over them, meaning you could get the direction the fish was trying to head in.

Price for the electronics comes to a few quid including the signal conditioning to let the tiny change in light level make a big difference to the ADC output.

Alternatively you could use a whole bunch of fibre-optic strain gauges as the harness.

Sensors would be tiny and hidden from the fish. The fish would be in a harness so it'd not end up looking behind itself or anything.

Then get that same computer to get the blimp to move in that direction. Speed would be controlled by figuring out the speed that the fish would achieve in water (or measuring its forwards force directly if you used the strain gauge idea).
( uberdeity Wondered why Owen Lars knew of the term "hell" on, Fri 31 Oct 2008, 14:29, closed)

Simple
Paint the fish's nose with Tippex and put sensors on the outside of the bowl!
(daddypigsaw Geddinthebackothevan!, Fri 31 Oct 2008, 14:46, closed)

And whichever way it turns you know!
It'd be a good idea, and at face value far simpler than mine, but you'd need to know how fast it wanted to move as well- how'd you manage that?

And what sensor would you use to pick up the dot? A camera?
( uberdeity Wondered why Owen Lars knew of the term "hell" on, Fri 31 Oct 2008, 14:54, closed)

The thought I had
was to use strain gauges to harness him in place- in other words, one on each side, one on top, one on bottom, and one each front and rear. They would be the tethers that held him stationary within the bowl, but would tell the controller which way he was trying to go.

The problem? Carrying enough water, a pump, a filter and a computer on a blimp. The RC car idea is more practical, but even so...

The other aspect, of course, is the cost of this. I can't imagine it would be cheap, and I don't have a few grand to shell out on the stuff.
( The Resident Loon Not a demographic. Do not measure., Fri 31 Oct 2008, 15:33, closed)

Sounds overengineered to me.
Easier to tether the little blighter to a simple eight direction, weakly sprung 'joystick' type affair mounted to the top of the tank. Analogue if you like for the speed aspect. Saves making our piscine friend into a tiny version of Jones from Johnny Mnemonic.
(Prof. Moriarty put his foot in it on, Fri 31 Oct 2008, 21:43, closed)