Robotic Sushi - FAO Davy

[Guest]

Hi Dave,

Just following up on our chat at Merstham regarding programming PIC's, and control, of a RC robotic fish.

I've been doing some hunting around and came up with the following links-

A chip which interfaces with a OOPIC to control speed and position of upto eight servos (need four servos for a fish)-

http://www.totalrobots.com/pdfs.....01.pdf

To obtain a realistic undualting wave swimming motion requires the servo's to be positioned with reasonable accuracy, but also for them to be timed from position A-B. The above chip achieves this with minimal load on the microprocessor.

A coprocessor to deal with incoming PWM signals from the receiver. Takes the weight off the PIC chip- not sure if this is necessary or not-

http://www.awce.com/pak7.htm

OOPIC chips. These are PIC's that used object oriented programming. They can be programmed in BASIC, C or Java.
As such they look a lot simpler than conventional PIC's to program, and are well developed for robotics applications-

http://www.totalrobots.com/oopic.htm

These are the servo's I thought most suitable for the fish, as they're designed for robotics.

http://www.lynxmotion.com/Product.as...&CategoryID=91

More information on robotic fish here-

http://privatewww.essex.ac.uk/~jliua/

Lots of downloadable PDF files there, with plenty of interesting information.
I've seen these fish in action at the London Aquarium, and they're very impressive indeed.

Andy Lawrence

[scott t]

Interesting subject. I was wondering

Interesting subject. I was wondering how a swimming tail action might be produced after looking at the hobie mirage drive.

Hobie mirage drive.
I will have to study this fish project a little closer.

-Scott

[Guest]

Hi Scott.

If all you require

Hi Scott.

If all you require is a fish swimming motion, then look specifically at this article-

http://privatewww.essex.ac.uk/~jliua/paper/iros05formatpublish.pdf

It shows how to build a mechanical tail mechanism, powered from a single servo (you could use a crank or scotch yoke instead, if you have another means of turning the fish). The rest of the mechanism is pure leverage/linkage.

I've built a mock-up of this tail in cardboard, and it works well.

The disadvantage of this arrangement, is it lacks the manoverability of the servo jointed fish, so sharp turn-in-it's-own-length manoveres aren't possible.

Andy

[anonymous]

Hi Andy,

Just got back from

Hi Andy,

Just got back from holiday - sorry for delay in reply.

I will have a look at those interesting links - then come back.

David

[anonymous]

Hi Andy,

Fascinating paper - the

Hi Andy,

Fascinating paper - the MT fish model looks very practical.

I'm still trying to get my "head round" how exactly it does turns?

A back and forth swimming motion at 2Hz - Yes.

A varying "neutral" position to give the turn???

I'm sure a single PIC could do all the processing.

Having to reverse the main drive motor at 2Hz will only work for small motors.

Is there no way of bending the whole tail fin by servo to get the turns?? (Something similiar in Norbert Bruggens book but maybe this does not give a realistic swimming motion.)

Thanks for all the fish!

David

[anonymous]

Hi Scott,

I've just been taking

Hi Scott,

I've just been taking a look at the Hobie drive and isn't it clever!

Maybe the days of props are numbered.

David

[Guest]

Hi Andy,

Fascinating paper - the

Hi Andy,

Fascinating paper - the MT fish model looks very practical.

I'm still trying to get my "head round" how exactly it does turns?

It turns by cutting the stroke in half. The whole tail is controlled by a single servo, so the microprocessor just commands the servo to only swing from neutral to left/right as appropriate to instigate a turn.

I've made a mock-up of the MT1 tail in card, and it works fine.

However, it lacks the manoverability of the 4-servo jointed fish.

Lots of videos to download, showing all the types of fish, including the MT1-

http://privatewww.essex.ac.uk/~jliua/videogal2.htm

http://privatewww.essex.ac.uk/~jliua/videogal.htm

Still studying that oopic book.........brainsache.

Andy

[anonymous]

Hi Andy,

They move beautifully. I

[color=#000000]Hi Andy,

They move beautifully. I see the attraction.

A simple MT type of R/C fish may not be too difficult]

[Guest]

Well if you can come

Well if you can come with something in PIC form, Dave, then I can certainly rustle up the hardware.

I have one of those little RC Sharks like Brian Alps.

I was surprised at the radio range he got from his- mine was woeful, just a few feet (along not down), and the signal was lost. I booted the transmitter voltage upto 12 volts- crude but effective.

The batteries onboard are hopeless, and I just don't think it's powerful enough. The turning circle is too wide, it's tiny and it's a tricky little critter to trim.

So quite a few moans, and I think it can easily be improved upon. I guess you can't complain too loudly for the money these things sell for.

The advantage with the MT1 design, is that it has a full body undulating motion design, which makes it a better swimmer than the little R/C shark which just flaps it's tail from side to side courtesy of a scotch yoke.



Edited By Sub culture on 1150304157

[anonymous]

Hi Andy,

The marking got boring!

In

Hi Andy,

The marking got boring!

In a funny way I'm glad that the shark is not too good. Ballasting must be a problem because it is a flexible bag and air filled - an interesting point from the Essex Uni paper.

Have you thought about contacting Essex for a simpler device? They seem to have gone for an autonomous robot which makes it complicated. We still want some sticks to twiddle!

I will have a go at a simple PIC version once I have finished my marking and sent the latest Pinger order to Australia. (Thank you Australia!!)

In the meantime my head is still hurting a little from the Essex paper but am I right in thinking that a 2Hz sine wave with a fixed cut-off point on each side for turns would do it?
No more variables needed?

Now - more marking.

David

[Guest]

Well it would be handy

Well it would be handy to have a swimming motion on turns, but that could always be added later on.

Regarding autonomous control, absolutely we want sticks to twiddle. However the research is equally applicable to an RC fish aswell as a standalone unit- instead of sonar and infra red controlling the direction of the fish, we want PWM from the receiver dictating the direction.

You can thank me later for the Aussie order, I put that chap onto you, after he posted a picture of him fishing about in a lake with a big stick looking for a lost sub.

Do I get commission?

Andy



Edited By Sub culture on 1150364609

[anonymous]

Hi Andy,

Many thanks for the

Hi Andy,

Many thanks for the Aussie order. My life's work is to prevent model submariners from having to poke sticks into muddy ponds! Commission will be a prototype Fish PIC (PISK?) to test in a few weeks time
Steering will form part of it from the start.

On another matter I'm heading off to see the big subs tomorrow - an open day to see the Astute and Ambush attack subs under construction at Barrow. No cameras allowed but I will try to write a piece.

David

PS I must also thank you for your suggestion to use polycarbonate stern fins on my R class - they ARE near indestructible.

[Guest]

Wow, wish I was going

Wow, wish I was going too!

Still I have a sub run to, erm, run.

A programmed PIC will be more than sufficient payment.

Well I had better start building a fish body- I have all the materials to hand. I have some slightly different ideas to the essex boys, the fish will be made out of composite epoxy glass sheet and foam- light, tough and bouyant!

Andy

[anonymous]

Just a little progress report.

Just a little progress report.
Going well. The servo action for "full ahead" is working well. (Even with a standard servo - which surprises me - I didn't think it could cope with rapid direction change.)

Now sorting out the turns. (These will be input by switches initially rather than PWM - packing the code in for PWM may take some time!)

Andy, I'm not sure that I have your address. If you could email me privately I will post the prototype to you, hopefully in the next few days. How is the tail linkage coming on?

David

[Guest]

Hi Dave,

I've been playing about

[color=#000000]Hi Dave,

I've been playing about with different materials- making it lighter than the version the Essex bods built.

Using epoxy glass sheet and foam composite- self bouyant strong.

On one note, i think the fish will need a high torque servo. Thankfully these are quite cheap these days.

I've tried emailing in the past, but your email bounces.

My email is]