Robector

[Jaxom]

Since my orginal post has started a thread that has gone beyond he scope of the forum i posted it in, any other posts pertaining to this project I'm moving here. The oringal post was the thread titled "Tower of Babel", in the suggestions to the editors of Robot magazine forum. Basic recap, I'm inquiring about comparisions of the various languages people use to program bots with. Along with example block diagrams as a side bar. A few of the other members inquired as to exactilly what I was trying to accomplish. This is where I went into length describing the "Robector" an automous metal detecting robot.

For a platform, I'm going to start saving my pennies for a MMP-40, made by The Machine Lab. If you go to thier home page at http://www.themachinelab.com/ and look in the upper right hand corner, I'm highly interested in that arm in that picture. I've written them asking for any information they can offer about that arm. If anyone has other ideas for a simlar arm, I'd appericate any info offered.

The detector I'm interested in is a Minelab Explorer II. This can be found at
http://www.minelab.com/ This is a good general purpose detector that can find coins, jewlery, and if have enough money left over form building the robot, I would like to take it to arizona or california and do some nugget detecting.

Well, that should be a good start for moving this thread, best be getting to bed, i don't have a robotic lawn mower....yet, and have to "get 'er done" before we get hit with several days of rain coming.

Jax

[HarryBotter]

Did any of my musings on the other thread make any sense? If so, I'd guess that $5000 for a robot platform is overkill. You can do quite a bit with a couple of wheel chair motors, a motor driver and a microprocessor.

Have you asked Minelab if they would sell just the coil and control box?

Or maybe you just have deep pockets. :)

Cheers, Harry

[Smokeypipeman]

I think the direction you are looking at will cause you a lot of problems.

What I would do(Now I am also thinking about doing it myself) is to make a wooden frame using plastic bolts. The wheels sit wide, about 1.5 meters square. The electronics sit a platform about 1++ meters above the wheels. The detector sits under the platform in the gap between the wheels. The sweeping motion would be conducted using something like a belt system, so it sweeps left to right.

The only real problem I can see with it is motor noise distorting the magnetic signal of the detector.

If you were to use the arm, the problems that would arise are bounce and stability on sweep. So you would have a lot of stuffing around. And also the balance of the platform on uneven ground(lots of nose diving which may break the detector in a field test)

[Jaxom]

Ahhh, it's good to be back! Dummy me had a bad case of not having my virus software updated and then getting some cad data from a friend that also provides home pc repair! (mainly removing viruses and stuff from newbie computer users)

No, i don't have deep pockets. In fact, they're rather shallow, as I'm unemployeed at the moment. Although i do start school on the 25th of sept, to achieve my welding cert.

One idea I had about moving the loop, rather then sweeping it along with an arm tyring to mimic a person, as to have some sort of printer head arrangement. The concept being that the bot would move into place, then a controller would move the loop back and forth over the spot then move the whole "carriage" forward to achieve the over-lap. Once the loop has scanned the area it's over going in one direction a second carriage would then move the front to back, instead of from side to side. This would achieve the double scan at a right angle as I've previously mentioned.

Issues.... To initialize the loop it has to be raised with it's edge side facing down to ground balance and then touch the ground, on edge to complete balencing just to get going. While scanning, to acccuratly pin point you raise the loop up and down by 6-8".

UGH!! if only ground penetrating radar was more sensitive and show details of things that are shallower then 10" deep! This could be soooo easy! Having read the story about an erthernet controlled bot, and want to control my in a simalar fashion, I went to that companies website. And what do ya know! Not only do they sell the parts for that, they also have simple schematics for a metal detector! I could actually go to the extent of building this, rather the potentially wrecking a $1500 unit! Only if... I could find plans/schematics for ground penetrating radar that can "see" that first 12" deep.

Jaxx

[HarryBotter]

Ahhh, it's good to be back!

Good to see you back.

One idea I had about moving the loop, rather then sweeping it along with an arm tyring to mimic a person, as to have some sort of printer head arrangement. The concept being that the bot would move into place, then a controller would move the loop back and forth over the spot then move the whole "carriage" forward to achieve the over-lap.

Sounds like a good idea.

Once the loop has scanned the area it's over going in one direction a second carriage would then move the front to back, instead of from side to side. This would achieve the double scan at a right angle as I've previously mentioned.

This sounds like a needless complication. Why not just have the bot drive the grid a second time at right angles to the first. Much easier to implement.

Issues.... To initialize the loop it has to be raised with it's edge side facing down to ground balance and then touch the ground, on edge to complete balencing just to get going. While scanning, to acccuratly pin point you raise the loop up and down by 6-8".

Couldn't this be done manually before the start of every session? KISS is a great principle for first prototypes. Get something that works first and then make improvements rather than trying for the stars in your first shot.

Cheers, Harry

[Jaxom]

You're absolutly right Harry! I was giving the KISS meathod a thought about the whole project. Yes i could get away without doing the pin-pointing, and I could possibly even skip the ground balencing proceedure as well.

After I posted last night, I was thinking about that bot that the magazine did a story about about an ethernet controlled bot. The one with the linksys router mounted ontop... That same company that offers those parts also has diagrams and a C program for a metal detector! A very simple one mind you but it would still work.

As far as the arm is concerned, instead of what we would normally concider an arm it would just be a bar or rod sticking out the front end, far enough away motor/metal won't interfer with the loop. Still using the printer concept of moving the arm back and forth.

The metal detecting circut i talked about would need to be modified a bit. One of the better models of gold detectors runs at a frequency of 71khz. Instead of sound the small circut uses leds to display metal being presant, intead of that i would need to send that data to my lap top.

But I think now it's time to start making some baby steps. And planing out how to proceed with this. Admittedly the MMP-40 is WAY out of budget, i was wishful thinking there. But thats not to say I couldn't build something very simlar. I like the wide treads and the ability travel over rough terrain.

As far as a first bot to build, i'm really leaning towards a surveyor bot. Only slightly pricey, but it has many of the features that i'm looking for. Would that be a good bot to start out with?

Jaxx

[HarryBotter]

Congratulations Jaxx, you're starting to sound like a robotist! Start small and start researching some of the things that may challenge you in the bigger project.

The Surveyer would make a great project, I've looked at getting one myself.

I'd still look at wheelchair motors and tires to push my contraption around, they're much more accurate than tracks if you have to know where you've been and where you're going.Best way to work that out is to try it with small prototypes.

Cheers, Harry

[Jaxom]

Darn it all, the Surveyor is pre-assembled! Just did some checking around, looks my other options are lynxmotion's 4wd1 or the Stomper. Again, both have many of the features that I'm looking for.

While I'm saving for that, I'd like some suggestions on some good robotics books. I've already bought one of those Dummy's guides and I downloaded an Ebook on to build and program a bo-bot in java.

Speaking of languages, I'm still very confused about them. Some seem to widely accepted, while other are propritory to specific microprocessors. At this point it's hard to say just exactilly what sensors I will need, although by now I have a good idea. It's just a matter of which brand and how to intergrate it into this project. But if you all had to build a bot using one language, which has the most accesories/sensors avaible for it?

jaxx

[HarryBotter]

Hi Jaxx,

You're getting to the point where personal preferences kick in and they vary all over the place.

Three beginner's books that I liked are:

Mobile Robots - Inspiration to Implementation by Jones, Flynn and Seiger
Intermediate Robot building by Cook
Robot Builder's Bonanza by McComb

I think Lnnxmotion is one of the best sources of beginner type robotic platforms around. Good, products, good support, good prices.

Languages and sensors are not tied together. All a sensor needs are some pins on a microprocessor and some instructions to on how to handle the sensor output. You've got to decide what language best suits your needs and how to best implement it.

I like C and have made that my main focus however, I'm also quite comfortable with most forms of basic.

The ball's really in your court now. Check out lots of other projects on the net to see what they're doing. Read magazines and books from the library, soon it'll all come together. Good luck!

[wrighthobbies]

Jaxx,

Languages are often a matter of personal choice and cost. You can download most compilers and try out the language without buying the hardware.

If you don't have much programming experience, then some flavor of basic may be a good place to start. PBasic for Stamps is simple and pretty easy to understand. You can also download Bascom for AVR chips and try it out. Bascom is more capable (hence more complicated) and has built in support for a lot of peripherals.

On the C side, Harry mentioned WinAVR, the GNU C compiler for Avr that is also free.

If you don't mind spending some money, there are several really good compilers available for PIC and AVR.

So start out by writing some programs without ever buying a chip.