Go Down

Topic: Robot mower navigation (Read 2466 times) previous topic - next topic

rbtgeek

I am building a robot mower so I am interested in your ideas. I am thinking of using a digital compass for navigating in rows while mowing. This might be accurate enough for a small yard. I have been reading about triangulation with ultrasonic signals and wondered if anyone has actually tried this.

CrossRoads

What is your plan for creating boundaries? Bury a wire like a dog wireless fence system? Could probably use the collar then as the boundary detector too. Some have 2-level detection - beep when close, then shock creation at the boundary.
Using compass to tell it when to stop making a 180 degree turn after stopping at a boundary might work.
Have to keep in mind that compass will lead or lag the actual turn.
Ultrasonic - would that need to be pretty directional to be able to be received outdoors?  I don't know how quickly it dissipates with distance if omnidirectional.
Designing & building electrical circuits for over 25 years.  Screw Shield for Mega/Due/Uno,  Bobuino with ATMega1284P, & other '328P & '1284P creations & offerings at  my website.

rbtgeek

Yes I plan to use buried wire from dog fence or the harbor freight wire tracer.

I have no experience using ultrasonic triangulation so I hope to hear from someone who has. I haven't used ultrasonic outdoors but some range finders are made for outdoors.

PaulS

Quote
I have no experience using ultrasonic triangulation so I hope to hear from someone who has. I haven't used ultrasonic outdoors but some range finders are made for outdoors.

To use triangulation, the signal has to bounce off of hard surfaces that form the boundary. That hardly seems like the scenario that will be present in your yard.

Unless you are trying to minimize the time needed to mow the yard, a roomba or pool cleaner process would work - mow in some direction until you can't go any farther; turn some random amount, and mow some more. Sooner or later, the whole yard is mowed. Of course, it might look like a drunk did it, but the grass will be short, and you can drink while the yard is being mowed.

zoomkat


I am building a robot mower so I am interested in your ideas. I am thinking of using a digital compass for navigating in rows while mowing. This might be accurate enough for a small yard. I have been reading about triangulation with ultrasonic signals and wondered if anyone has actually tried this.


Previous lawn mower discussions that might provide helpful info.

https://www.google.com/search?hl=en&as_q=mower&as_epq=&as_oq=&as_eq=&as_nlo=&as_nhi=&lr=&cr=&as_qdr=all&as_sitesearch=http%3A%2F%2Farduino.cc%2Fforum%2F&as_occt=any&safe=images&tbs=&as_filetype=&as_rights=
Google forum search: Use Google Advanced Search and use Http://forum.arduino.cc/index in the "site or domain:" box.

rbtgeek

#5
Feb 16, 2013, 10:54 pm Last Edit: Feb 17, 2013, 12:34 am by rbtgeek Reason: 1

Quote
I have no experience using ultrasonic triangulation so I hope to hear from someone who has. I haven't used ultrasonic outdoors but some range finders are made for outdoors.

To use triangulation, the signal has to bounce off of hard surfaces that form the boundary. That hardly seems like the scenario that will be present in your yard.

Unless you are trying to minimize the time needed to mow the yard, a roomba or pool cleaner process would work - mow in some direction until you can't go any farther; turn some random amount, and mow some more. Sooner or later, the whole yard is mowed. Of course, it might look like a drunk did it, but the grass will be short, and you can drink while the yard is being mowed.


Yes of course the random idea will work but it is inefficient and not as fun as planning some smarter navigation.

For the triangulation I should be able to place 3 transmitters in the yard and use a receiver on the mower and measure time of flight from the transmitters to determine current location. I have never tried this so I'm unsure of the issues I will face.

zoomkat

Quote
For the triangulation I should be able to place 3 transmitters in the yard and use a receiver on the mower and measure time of flight from the transmitters to determine current location. I have never tried this so I'm of the issues I will face.


Get that to work and you will be an instant forum guru!
Google forum search: Use Google Advanced Search and use Http://forum.arduino.cc/index in the "site or domain:" box.

CrossRoads

http://en.wikipedia.org/wiki/Ultrasound
Grrr, lost my message twice now!
Boiled down - have each transmitter determine distance from reflector equipped mower, send the distance to the mower (via RF I would think). Mower would be along an arc of distance from each transmitter, do the math to determine where the 3 arcs intersected.
That might be doable.
Designing & building electrical circuits for over 25 years.  Screw Shield for Mega/Due/Uno,  Bobuino with ATMega1284P, & other '328P & '1284P creations & offerings at  my website.

rbtgeek


http://en.wikipedia.org/wiki/Ultrasound
Grrr, lost my message twice now!
Boiled down - have each transmitter determine distance from reflector equipped mower, send the distance to the mower (via RF I would think). Mower would be along an arc of distance from each transmitter, do the math to determine where the 3 arcs intersected.
That might be doable.

Do you suggest doing the processing elsewhere because you don't think the arduino can do it? I was thinking the transmitters would be dumb transmitters only and not transceivers. The mower would have a receiver only. I am open to suggestions however.

CrossRoads

I am thinking have the transmitters do the distance calculation because they are not moving.
If you have the mower do it, then you are making it do 4 times the calculating (along with the actual mower control).
Read the article - all the descriptions are of a source measuring the time from transmit to receive to determine distance.
What you are describing is more GPS-like - the transmitters would need to send a time of transmission, the poor receiver would have to pick out one of three signals, interpret the message, etc. So not just dumb transmitters, but transmitters of a time message encoded in an audio wave.
If you break up the processing, I think you will have more success.
Designing & building electrical circuits for over 25 years.  Screw Shield for Mega/Due/Uno,  Bobuino with ATMega1284P, & other '328P & '1284P creations & offerings at  my website.

michinyon

Interesting spray from James Dyson today,  claiming that robot vacuum cleaners are "rubbish",
which is why he hasn't made one.
I don't know if mowers have better navigation than vacuum cleaners do.

Tomalo

You could map out your lawn and have the robot go a set path, so long as it doesn't have any interference.

rbtgeek


You could map out your lawn and have the robot go a set path, so long as it doesn't have any interference.


Yeah but not sure how to map it out. Too bad GPS is not accurate enough.

Go Up