gps guided lawnmower

My new project is going to be an automated lawnmower. I think i will use the mega 2560. My goal is to modify an electric lawnmower to cut an area of grass 50' by 50'. I thought gps would be the perfect choice, would it be accuracy enough to make a turn and cut a new path perfectly aligned with the last path? Any suggestion for additional sensors to improve accuracy.

1 Like

Electric as in battery powered? If not then handling the power cable will be an issue.

I am doubtful about using gps for this as I think making accurate turns will be difficult. Do many of the commercial robot mowers use gps? Why not try it out using a handheld gps and walking round with a mower.

I have heard of mowers following a wire buried under the lawn.
You will need to be sure the machine cannot runover pets, children and your own toes.

i have read about the berried wire method. I would like to create some thing a little more simple to set up.

as far as safety precautions all though it will be tedious and time consuming. I think there are enough general sensors available to guarantee safety.

The problem I am trying to work out as you stated is the lack of accuracy with gps. I am thinking with a base station (witch I am still reading about it will improve accuracy to with in centimeters witch is better than 3 feet. another option I am contemplating is triangulation with radar. I have heard that some model airplane hobbyist are programming there plans to take off fly a pattern and land autonomously but i have yet to find anything that says how they are doing it.

edit: yes the mower is battery operated 24v
edit 2: with regards to triangulation with radar I wonder if I do not get the reading simultaneously would it skew the position.

Getting position within centimeters as you’re describing will be extremely difficult. As far as I know, the commonly available commercial automatic lawn mowers work by patrolling within the perimeter on a random pattern with the idea being that eventually their path will cover every part of it - they don’t work in neat lines as you would if you were cutting the lawn manually. If you wanted your mower to follow a set pattern, I think you would need some sort of beacon system so that the mower can aim towards the far side of the lawn, move sideways by the cut width, and then aim for a beacon at the far end of the next row etc. I expect this would need some way to keep track of where the mower was and coordinate that with turning the beacons on and off as required.

As a first step, a simple remote controlled lawnmower would be massively easier to build and a lot more fun than mowing the lawn yourself.

PeterH:
Getting position within centimeters as you're describing will be extremely difficult. As far as I know, the commonly available commercial automatic lawn mowers work by patrolling within the perimeter on a random pattern with the idea being that eventually their path will cover every part of it - they don't work in neat lines as you would if you were cutting the lawn manually. If you wanted your mower to follow a set pattern, I think you would need some sort of beacon system so that the mower can aim towards the far side of the lawn, move sideways by the cut width, and then aim for a beacon at the far end of the next row etc. I expect this would need some way to keep track of where the mower was and coordinate that with turning the beacons on and off as required.

As a first step, a simple remote controlled lawnmower would be massively easier to build and a lot more fun than mowing the lawn yourself.

You know, if I ever have a lawn again, I am going to build a remote controlled lawnmower. This sounds like an excellent idea.

But yes I've seen graduate level theses on figuring out how to do this type of thing, it is a non-trivial problem to do reliably. Aren't there some forms of Local positioning systems that use beacons, similar to GPS, to then calculate your location relative to the beacons?

You know, if I ever have a lawn again, I am going to build a remote controlled lawnmower concrete it over and paint it green. This sounds like an excellent idea.

As a first step, a simple remote controlled lawnmower would be massively easier to build and a lot more fun than mowing the lawn yourself.

agreed! it would be easier. But not as rewarding.

I think you would need some sort of beacon system so that the mower can aim towards the far side of the lawn, move sideways by the cut width, and then aim for a beacon at the far end of the next row etc. I expect this would need some way to keep track of where the mower was and coordinate that with turning the beacons on and off as required.

anyone have any thoughts on using radar for triangulation? lets say I put a reflector on each corner. would I get faulty readings due to angle variations? Would it be completely out of the question? I really do not seen much info on the idea.

another thought using cameras, mounted around the the mowing area and then use photo recognition software. and http://www.sikuli.org/ to automate a remote control.

Previous lawnmower discussions.

https://www.google.com/search?as_q=lawnmower&as_epq=&as_oq=&as_eq=&as_nlo=&as_nhi=&lr=&cr=&as_qdr=all&as_sitesearch=http%3A%2F%2Fforum.arduino.cc%2Findex&as_occt=any&safe=images&tbs=&as_filetype=&as_rights=

kennagle:
anyone have any thoughts on using radar for triangulation?

Radar isn't really practical for DIY because it involves dealing with speed-of-light signal propagation and some pretty clever signal analysis. What would IMO be more feasible is an optical beacon that a sensor on the mower could detect and do crude direction finding on to discover which direction it was in and then head towards it, then have some other means for the mower to discover when it had reached the edge of the mowing area and turn/move into position to mow the next line, switching the old beacon off and the next beacon on. It would be difficult and fiddly to set up and involve laying beacons and wiring and who knows what else around the outside of the lawn, but it all seems generally feasible for a small area. But before you consider automating it, just designing and building an autonomous mower would be a considerable challenge in its own right. I know that if I'd put all the time and money in to making that I wouldn't want to wait until I had got the SLAM problem solved before I saw it working - a plain old RC system would be all you needed to get it trundling around the lawn causing havoc while you enjoyed a beer.

Differential GPS Differential GPS - Wikipedia will give you almost the accuracy you need. I'll wager it isn't cheap though!

wildbill:
Differential GPS Differential GPS - Wikipedia will give you almost the accuracy you need. I'll wager it isn't cheap though!

Exactly what I was going to say, you beat me to it.

Location based on beacons seems feasible to me. Think about the Nintendo Wii remote sensor bar. Mounting three of these (or something similar) in a triangle on the mower, and placing at least three beacons in the yard, should allow location finding by getting a bearing (not time-of-flight) on each one plus a little math. Each beacon will need to transmit a different signal though so you can distinguish them. Also, IR is not the best choice outdoors because sunlight interferes with it. Using a UV wavelength not found in nature may be an option, but power levels need to be low enough to not cause retina damage to bystanders.

Sorry if this is an old topic, but I am also in the process of building an automatic lawn mower. Currently I am still under R/C control but I have plans to make it autonomous.

Has anyone thought about putting a kinect sensor on the mower? It would give you vision and can also double as object detection/avoidance sensors. Im wondering if it would even be able to visually detect the the cut line from the previous path and line the mower up to mow right next to it for a straight line.

You could use GPS just to create a boundary around the yard and let the kinect do the rest.

robovoyager:
Has anyone thought about putting a kinect sensor on the mower?

You might Google "Kinect" and "Arduino" to see what others have done, to see if this is feasible.

Years ago I spent some time considering gps, diy differential gps and automatic lawn mowers.

What I learned is that both types of gps do not refresh fast enough to be acceptable for lawn mowing alone.

Here is the path I will go down when time permits:
1: DIY differential GPS for perimeter control (property lines and flower beds)
2: sonic object detection (stop it from going over objects in the way and to force it to slow down/stop till the differential GPS system can recognize the perimeter)
3: electric compass to maintain direction and orientation for turning

Many hours spent going down this path looking at so many options. It is great to find a group/blog that is discussing this topic and help this concept move along!

rcharper:
What I learned is that both types of gps do not refresh fast enough to be acceptable for lawn mowing alone.

GPS receivers have improved since then. You can now get receivers with 4Hz update rates. Fast enough to fly a quadcopter or fixed wing drone, fast enough to mow a lawn.

Also, take care mounting your magnetometer. Lawn mowers usually have magneto ignition systems, which is a big magnet spinning around on the flywheel.

you could have two beacons, place them at adjacent corners of the plot, the robot finds one beacon, and lines up on the 2nd beacon, but before moving to the 2nd beacon, it picks up and moves the first beacon a preset distance over to the next return position, then the robot goes to the 2nd beacon and repeats the procedure a set number of times

I'd go with a Toro Self-Step ( i think thats what its called) ive had mine for 5 years and it runs great, the good thing about it is that the speed adjusts to how fast you walk so you don't have to worry about adjusting the speed and all that.