my autonomous lawn mower

[epicjr77]

http://youtube.com/watch?v=bIHgdfiNxUg

[airman00]

very very nice job!

Any more documentation?

[Admin]

nice video.

but . . . the grass didn't get cut in the video?

also, what software did you use for coverage testing?

[vidam]

Was it a push lawn mower originally. How did you modify the powertrain so that the computer could turn the wheels rather than being pushed by a person?

[Private Reid]

I like!

Although a little inefficient...

The robot doesn't know if it has already cut that part of grass,

and it doesn't know what it's cutting! For all it cares it could be cutting your prize vegetable patch! or trying to cut a concrete patio...

But apart, I like!

[inkel]

I think this is a fun idea.
but i dont think it can be used for real.

[izua]

I think it can. Solar powered beacons can be installed in the garden, and the robot can tag them as he passes. This way, some kind of primary mapping can be done.

[Private Reid]

Yes thats correct, use solar powered (with battery backup) "posts" to mark out a perimeter of the garden.

Then the bot can cut anywhere in the "fence"

If you get what i mean

Still, a little dangerous.

[praig]

I would love to someday build a lawn mowing system and so saw this post. I think its great! So in my ignorance I have to ask, incorporating a gps system seems to be a quick solution to so many of these "mapping issues". Is it that hard to do that so few people tackle it?

[airman00]

GPS is not accurate enough . There will be like a +- 5 ft error. You need sensors accurate to like an inch.

[praig]

If you want to cut 100% of the grass or the yard is small, your right its not worth it at all. I'm thinking larger scale though (acre or more) and effort reduction (70-80% grass gets cut).

I dunno, maybe it better to hire some workers.

[steferfootballdude13]

GPS is not accurate enough . There will be like a +- 5 ft error. You need sensors accurate to like an inch.

You can get GPS accurate to within about 2 inches, however they are very expensive, i don't think they work (well) while moving, and are not a good option for this anyways i would think.

Why not just program the robot for it to cover the area instead of having to make it autonomous? It (after programming) would be more efficient as you could manually map out where it would drive. (aside from having to manually map every route you want it to take)

[Admin]

You can get GPS accurate to within about 2 inches

you mean feet?

[steferfootballdude13]

You can get GPS accurate to within about 2 inches

you mean feet?

No, i met a guy from the USGS that showed me one that was accurate up to about 2-4 in. He was a surveyor of some sort.

[steferfootballdude13]

http://www.racelogic.co.uk/?show=VBOX-Products-VBOX_III_RTK&gclid=CMOv44vnnZUCFQMIswodYSu_jQ
"you obtain a phenomenal ±1cm position accuracy which is accurate enough for a range of diverse and demanding tests."
That, if true, is amazing.
I also like how they use a caliper in the first picture next to it.

[Spoil9]

nice, but it said they use "the Russian Glonass range of satellites as secondary observations"
never heard of them, but I bet they don't help us much here in the states.

[Admin]

but it said they use "the Russian Glonass range of satellites as secondary observations"

Thats what really surprised me. The US constellation is much more accurate that the Russian one . . . But it said it had its own 'differential Base Station', which I guess improves the accuracy.

I'd like to see this independently verified. Also, the price tag

[airman00]

I had this idea of making your own "GPS" network but small scale. It wouldn't really be global but rather it would be only in your yard or something. You have like three RF stations positioned in the yard and then you check how far you are from each station and calculate your position in the yard.

Anyone heard of this being done before?

[ArcMan]

I've heard of something similar, which I want to eventually play around with.  Using RF or IR coupled with ultrasonic to measure time-to-travel of the ultrasonic signal on 3 or more transponder stations to triangulate your position.  I'm not sure how well that would fare in the yard, though.

[colorclocks]

I think this is great, and I don't think it needs fixing to be the main lawn mower for someone's yard.  Of course you're going to have to trim more than you would if a conventional lawn mower were cutting the grass.  But I don't think people need to worry if a robot mower doesn't cut everything on the first pass, or if it makes several passes over grass that has already been cut.  In fact, I think these things are advantages for robot lawn mowers, generally.  Here's why:

It just kills me that commercial robot lawn mowers currently cost more than the average conventional mower, when they could easily cost less.  (Of course folks will pay more, because robot mowers are an improvement over manual mowers.  But I'm talking only about what it costs to make a machine that will just do the job.)

Robot mowers should cost less because they have an easier job to do.  They don't have to cut quickly, and they don't have to cut everything well in a single pass.  A conventional mower has to do these things because the person pushing it wants to get the job done quickly and with minimal effort.  But a robot can mosey over the lawn all week.  So a robot mower can be much less powerful, and it doesn't need a sophisticated control system to concentrate effort on uncut grass.  Also, a robot mower isn't going to be pushed hard, reversed quickly,  banged into trees, etc., so it doesn't need to be as solidly constructed as a conventional mower.  It can be light, weak, stupid, slow, and cheap, and it will still get the job done.

I know this is true because I also built a robot lawn mower, and it's been the only mower cutting my back yard for a year and a half.  It's not light or slow (though the next version will be), but it's weak, stupid, and cheap.  And it proves that you don't have to do the job well all at once; you just have to prevent the grass from getting high over time.

Someday, someone will market a robot lawn mower for around $300, and everyone will buy one.  I think this mower represents a couple of steps in the right direction.  For another step, check out the Chinese mower on youtube:

[youtube]http://www.youtube.com/watch?v=8c5yFe2KEBg[/youtube]



That's not my mower, by the way.  Here's my ugly as sin contraption (shown at 8X normal speed so you can see how it covers the yard):

[youtube]http://www.youtube.com/watch?v=ah-p_C1o0yc[/youtube]

[Admin]

On that note, as colorclocks points out . . .

a solar powered robot lawn mower
http://www.google.com/search?q=Husqvarna+solar+power+lawn+motor
(props to whoever finds the price tag on it)

and a quick survey of robot lawnmowers:
http://www.robotics.uc.edu/papers/paper2000/lawnmower.PDF

[airman00]

Do something like this
[youtube]
JR774XxmLBs[/youtube]

[Admin]

Do something like this

so uhhhh . . . how do they cut the grass where the wires are implanted into the ground?

[Penth]

Do something like this

so uhhhh . . . how do they cut the grass where the wires are implanted into the ground?

They have a RC helicopter with a weedeater attachement.  Actually, as long as you stake them in close enough to the dirt, the mower blades shouldn't get near it, hopefully.

[airman00]

Do something like this

so uhhhh . . . how do they cut the grass where the wires are implanted into the ground?

they don't thats the point . I think if they use a Hall Effect sensor they will know when they are in proximity with the electrified wire. The wire acts as a "force field"

[colorclocks]

Do something like this

so uhhhh . . . how do they cut the grass where the wires are implanted into the ground?

they don't thats the point . I think if they use a Hall Effect sensor they will know when they are in proximity with the electrified wire. The wire acts as a "force field"

The easiest thing I know of is to drive the electrified wire at a relatively low frequency (~15 kHz, for example, where much of the radiated power is magnetic), and detect it with an L/C circuit tuned to that frequency.
All my mower has is an inductor and a capacitor in parallel, connected to a comparator, whose other input is held at some very low voltage (~25 mV in my case).  The comparator output is DC when you're away from the wire, and oscillates when you get within six inches or so.  If you feed this into a 555 timer, you can get a TTL low when you're away from the wire, and a TTL high when you're near it.  The parts for this kind of receiver cost less than a dollar.

Anyway, you have three choices about what to do with the wire:
1) stay away from it
2) bury it an inch or so underground
3) stake it carefully and wait a week or so for the grass to grow over it.

I chose 2, and that works pretty well, though burying is clearly harder than staking.

[JoeDodge]

Hello Guys
WOW!!!   Society of Robots.  I just discovered this great site. What a great place to see what other people are doing in their spare time.  Great ideas!!  Check out my lawn mower bot on Youtube.com    search for joestoys123      .............. joedodge