Slashdot Mirror
Building A Homebrew Robotic Lawnmower?
mmonkey writes "With the seemingly small amount of summer we get here in the UK, the last thing I want to be doing on a sunny day is mow the lawn. So I started thinking "surely a light-ish lawnmower could 'gain' a couple of motors, and suddenly be computer-controlled?". Then I started thinking about stuff like obstacle avoidance, optimum path planning, guidance system, how to get pretty-looking stripes, and I realised that it's actually a potentially complex (read: fun) thing to do. So, have any Slashdotters done this before? Did you modify an existing lawnmower or build a whole new one from scratch? What motors work best? For that matter, what type of mower works best? I know you can already get these, but that detracts from both my geek-drive and my wallet, both of which I'd prefer to keep as full as possible."
If you know what objects are fixed,such as pathways, bird feeders, what-not, you could build the controller from one of those old dump trucks from the 80's that let you pre-program a course by feet and angle of turn, etc. All you need to add is a bar attached to a kill switch for when the neighbors cat/kid/dog runs over to check it out.
If you are one in a million, then there are six thousand people who are just like you.
How about some duct tape job of roomba and a lawn mower ? You get obstacle avoidance and area coverage for free. You can even come up with interesting names like "Rower" or "Moomba" :^)
My first reaction was, "Well, you linked to what looks like a small business site, so either that site's going down or the hosting fees with bankrupt the company". But I digress.
that detracts from both my geek-drive and my wallet, both of which I'd prefer to keep as full as possible."
Well, I think your wallet's going to be drained either way. You need specialized components, software, etc for a completely automated solution. And even that's not going to be the end-all (corners, adjacent to fences, etc)
I would say start with a remote-controlled (as opposed to computer-controlled; mods, there is a difference) solution, see if you can rip apart some RC Cars, take their steering equipment out, see if you can interface to them using a RC Helicopter Remote or RC Airplane Remotes, connect up the servos, and perhaps sprinkle some detectors around your lawn.
Computer controlled would be difficult, to say the least. Perhaps even a Masters level thesis or a really good undergraduate senior project. Hell, if you can make it fairly cheap and efficient, you have your own business.
The safe way to do it is to have the thing run with very low power, and just nibble at the lawn, but do it all day. The big gas engine on the top of your push mower is really for the convenience of the operator, so that he can mow tall grass without stalling the thing or slowing down.
But if you mowed the grass yourself, one time, and then let loose your robot to simply maintain the height with a low power electric motor and some relatively safe blades, etc.... After all, it can stay out there all damn day!
eikimartinson.com
I read somewhere that in Scotland they use geese for the same purpose. You get a free trespasser alarm, since neither a goat or sheep will care about who enters the space, but a goose will attack first, then make a loud noise, and ask questions later.
Suppose you made a grill for the bottom of the mower deck similar to that which covers the blades of an electric razor. This would allow the blades of grass to reach the rotating blades while keeping fingers and toes safe. Perhaps finding one of those big furnace registers like you see in old houses would be the way to go.
Unknown host pong.
Yes, I would implement the following:
1. A _safe_ mowing method as you describe
2. The position measurement could be done with (ultra)sonic sensors and a transmitter on the mower. Doesn't work if you have to change garden often.
3. A power and communication cable. Can be cheap because of low power requirements because of 1.
4. A tower from where the cable goes to the mower and a mechanism to ensure that the cable doesn't get in the way.
Of course there is no obstacle avoidance etc., but I would start such a project in a modest size, not with all the 1000 features which _could_ be implemented!
As I'm doing such things also (Homemade microcontroller applications are everywhere in my home), I would say that it is far better to have a little thing working than big plans for a big thing but get nothing implemented.
If you really want to be different consider an autonomous swarm of mowing machines. The guts of a Roomba would be a good starting point! I'd like to see a self-organizing mesh network created by the mobile mowing agents.
Good luck - I'd love to see this when you're done!
First, get yourself a couple of Lego Mindstorms kits (so you have all the motors ans sensors you might require), and work up a useful collision-avoidin/path-cutting bot in your living room. maybe put it on a big sheet of paper, arm it with a felt-tip-pen, and tweak it's path-cutting algorithms like that.
Then, if you want to do more complex things - IR rangefinding, ultrasonics etc. strap a PalmPilot, Zaurus or some other PDS with IR on it and feed the midstorms controller unit with instructions from that.
Once you have it more-or-less foolproof (and you will probably want to run a wire round the maximum extents of your lawn and have a hall-effect or similar sensor pick up on it and kill the mower if it breaches that boundary) - then you can think about attaching a proper mower body and blade to it.
Then you'll probably want to port the whole thing to an embedded Linux u-Controller, and sell it for enormous profits.
I gots ta ding a ding dang my dang a long ling long
Seriously, if you go ahead with this, don't use a regular metal mower blade. Use something like a weed whacker--a nylon string. Coverage is far less and speed is less, but speed shouldn't matter in this application. So what if it has to make 4x as many passes...
Real programmers use "copy con program.exe"
I made a simulated prototype of a fast/simple algorithm, which was 100x (IIRC) faster than random wandering in my tests. A bit of information is here.
It requires that the robot know its position rather accurately, but if it's a hobby you could use differential GPS (which would add too much to the cost of a low-end commercial robot). You might look into localisation via wifi.
Pic Started mine about 11 years ago. The mechanical platform was a roboticized Toro 4hp mulcher using a permanent magnet motor driven backwards to generate power for two beefy wheel drive servos and the electronics. Fully autonomous. Narrow beam ultrasound sweeping the forward path for semi-coherent vision. No external environment markers used except where there aren't any objects to range off of for 20 odd feet. You walk it through the lawn once and it makes an internal map of the environment and the path you chose it to follow. Then, just plop it down and hit the start button next time. Works infinitely more efficiently than the commercial attempts at *cough* autonomous algorithms the crux of which is which way to turn after boffing into the perimeter wire or an obstacle. Rev. 2 is going to go battery-powered for safety and you'll have to "show' it where the charging station is. Wish I'd had money to take it commercial.