Robotic Lawnmower - its happening not just a dream

Forum 2005-2010 (read only) Software Syntax & Programs
1

Hello,

A good friend of mine who is a brilliant engineer has a lifestyle block just out of town. His ride on lawn mower just gave up the ghost, and while at the local lawnmower shop having a look around, he saw one of these robotic lawnmowers that drives around and docks to charge etc.
He gave me a call inquiring what I thought of the idea, effectiveness on his property etc, and the conversation ended up that we would build one.

His property is probably around 2000m2 in size, the part he wants to mow anyway. He is there for the long term, so having a system that is portable to another property isn't really required.

We thought about running (and burying) a inductive loop zig zagged throughout the area he wants to mow, so the mower effective is just a line follower. I however haven't used inductive type loops before, nor attempted to read one through soil. Does anyone know how effective this will be?

The mower he is going to design and build will most likely have a 1m cutting diameter, not one of these small ones that takes all week to mow a property.

I had toyed with the ideas:
GPS with set pattern.
Boundary wire and random pattern.
Boundary wire with random pattern with a couple of inductive loops for hard to drive areas which it would then follow.
Some sort of buried grid
etc
But out of all of these, the inductive loop buried slightly under the surface sounded the most feasible, esp when its not needed on any other properties.

Auto docking etc wont be a problem, I have an idea for that already. Its just the method of driving a known path and what is going to be the easiest or the most effective to do.

Has anyone detected an inductive loop buried before?

As I say in the title, this is not just a dream - it is going to happen in the next month or so as his lawnmower died and is very keen to pursue. His home workshop is better equipped than most engineering firms in town, so building something wont be an issue - he is very talented in the workshop also.

I may use just a 644P or maybe a 1284P and take advantage of the extra RAM - if its warranted. Else a standard 328 or something may be fine. It wont really matter, it will be on a custom made board etc.

Any input or ideas would be great, and any info on inductive loops would be appreciated.

Regards
James

2

MAKE magazine featured a remote controlled lawn mower in issue 22. (if you have haven't heard of it, its quarterly and issue 24 just recently came out.) The guy was planning on adding a gpsduino, mowing his yard and recording the path, and then using the gps plot to control it. His site is www.rediculouslygoodlooking.com (spelling error intentional) and its what got me interested in arduino.

His full schematic and code is posted here.

Its similar to what I am now working on, but my goals are set much lower for now. I was also liking the wire buried under the yard and having it drive basically like a roomba. I was thinking you could use an electric fence and a dog collar, but i need way more knowledge on interfacing the hardware before i could do something like that.

subscribe to the magazine, itll be the best 30-35 you ever spent.

3

As I say in the title, this is not just a dream - it is going to happen in the next month or so...

If it were a fairly simple/inexpensive thing to do, you would see a lot of completed autonomous lawnmower projects on the net. The mechanical side of things would be fairly simple, but the accurate locating of the mower position sufficient for decent lawn mowing is going to be a tough nut to crack. Start with the hardest part of the project and solve that before spending $$$ on the other parts.

4

Thanks to both of you.

RC robots are not what I am after - been there and done that before - but thanks.

RE expense and simplicity - expense isnt really an issue. Simplicity is based on what you compare it to - I can always learn something I dont currently understand.

My mate brought a $8000 ride on mower which lasted 500 hours. He was looking at a $17000 mower which had all the bells and whistles he requires, and is better quality etc. So if the cost of this robot mower is under $8000 to build and works, then it is a success.
As I say, money isnt an issue in this case, its just determining the best method to trial, build it small scale to start with and then build a proper well designed mower which incorportates the technology.

I have been reading about INDUCTIVE GUIDANCE in general, and from what I can tell, a 5W+ driver would be suitable, oscillating at something in the 30kHz range, running on 12V or 24V. Pump that down the wire and build a circuit on the mower which has 2 aerials to read the signal.
I havent found any example circuits yet however, nor have I put much thought into it, however it shouldnt be too hard I wouldnt think.

GPS options sounds good, however the accuracy is concerning, aswell as the storage of the path required etc. I dont think its worth pursuing for this project, especially when I havent played with a GPS into a microcontroller. Its just another thing to trial and learn, but the accuracy that is required compared to what I think I would realistically achieve from a GPS - doesnt scream success to me.

Keeping it simple, a line follower however with the slant of line following an inductive signal - I think is a winning solution. Building the driver and receiver, and the amount of power required etc is all going to have to be determined.

If anyone has ideas for this then I would love to hear. I am more than happy to share any learnings I find, schematics etc. This isnt a commercial venture.

Regards

5

An example:
http://www.agve.se/page/by_desc/inductive-guidance

6

As I say, money isnt an issue in this case

That opens a lot of possibilitys. Back in the 60's there was a race car builder named Mickey Thompson who had people comming by his shop and wanting him to build them a car to break some world speed record. The very first thing he would tell them was "Speed cost money. Now, with that understood, just how fast do you want to go?"

7

WanaGo:

I recently picked up a Friendly Robotics RL500 for cheap off of Craigslist (got it for $50.00 USD), with the intent of sometime in the future setting it up with a different controller to mow my back lawn.

Its original system was an inductive loop around the edge of the property, but instead used random path mowing for the mowing part. Its one of those things where the first few times out things look like crap (lots of missed spots), but over time, with it going out every day, on average the level of grass remains constant.

The biggest problem with an inductive loop or path is that if it breaks for some reason, finding the break may or may not be as easy as you think. So keep that in mind.

I was actually planning for my revamp (when I get around to it) to set things up so that the system can use some method of roughly knowing where it is, where its been, and where it hasn't, so that it can concentrate on mowing those areas that haven't been mowed, versus those that have. This would likely require some form of mapping and planning algorithm, and perhaps more than what can fit on an Arduino (though maybe a mega might work OK).

For your lot size, I would say a combo of edge inductive loop, with a GPS position mapping, plus some form of simple mapping/path planning - might work out best. I would also consider using some kind of engine on your mower; either as a hybrid solution (small genset coupled to a battery coupled to the motors), or use an engine for everything (perhaps using small hydraulic motors for the wheels, a PTO for a generator to power electronics, and some other system for the cutter blades).

Another option would be to "roboticize" a commercial ride-on mower; personally, for ease of interfacing, performance, and quality, a Walker mower would be my choice (they aren't cheap, though - even used).

Finally - make sure you put in place all manner of kill switches and safety systems (heartbeats, watchdog timers, etc); for the size of robot mower you are planning, you will want all this in place up-front (don't be tempted to bandaid this on at the end - like any good security system, it should be thought about, designed and built-in at the beginning).

Good luck with the project!

:slight_smile:

8

Thanks guys.

One of the sites I have found is this one:
http://www.philohome.com/sensors/filoguide.htm

It has a couple of example circuits for the small scale lego robot this guy built.