How About a M3DPMower?

Random or Off Topic
1

How about a mostly 3DP Mower?
One that is compatible with the OpenMower project?

OpenMower is based in Europe?
They have their own code, their own ESC motor controllers etc BUT, they don’t have a 3DP Mower with COTS (Commercial Off The Shelf) motors and body parts.

I’ve seen and purchased the plans for a different mower:

It is very clunky and uses an electronic dog fence wire surrounding the yard to be mowed. I like the wheels though.

I would like a smart mower. One that uses the GPS info from an RTK base station to mow my front and back yard intelligently. Like this one does:

What do you think, V1 Engineering? Can you come up with a generic mower that is Mostly 3D Printed and works with the Open Mower project? There are files to make the circuit boards involved, so use those measurements from the PCB files to make them fit in your M3DPrinted mower? And spec motors that can be controlled by their boards?

For the RTK base station, see the SparkFun RTK Base Station project that is very well documented here:
https://learn.sparkfun.com/tutorials/how-to-build-a-diy-gnss-reference-station/all

Granted, it’s not easy to design such a Mostly3DPOpenMower. Otherwise I would do it. But I thought I would ask the engineer that brought us such wonderful things like the Lowrider CNC, the MP3DP and the MPCNC.
And all of the wonderful people in the forums of V1 Engineering.

I am getting old and can’t mow my small yard.
And my SS income doesn’t allow me to easily pay someone to mow it.
In the years I am still working, I’d like to build such a mower. It will pay for itself. At $35 every two-weeks to pay someone to mow my yard, that’s about $455 per mowing season. One season should pay the $300 RTK station and the next mowing season would pay for the cost of making the mower.

I think this project fits as a new V1 Engineering project.

What do you think?

3 Likes
CCJ...The one tool you didn't think you needed but can't work without now
2

If you are not living in a HOA controlled area, why not plant flowers for the bees? Be a little lazy and do sth. for the environment. :smiley:
Even though I think the idea is cool, I am not able to design it. Sorry. :frowning:

1 Like
3

GPS isn’t accurate enough for a robot mower, that’s why all of them use the signal wire.

1 Like
4

A full yard of clover is actually nice looking.

1 Like
5

Rtk corrected GPS gives you 1cm accuracy (theoretically). But if you have trees or multi return paths (from the signals bouncing off of buildings), it can still give you 1m+ pops. You also need an imu and wheel encoders to include in your localization. The piksi GPS receiver is a decent all in one. GPS also doesn’t give you orientation. A digital compass can help, but electronics really mess with them. Once you move, you can use your movement to determine your heading.

Having a decent sensor will help you avoid running into things and can also be used for SLAM. But it needs significant processing.

The open mower project is collecting solutions in ROS, which is a good and a bad thing. It will be easier to get to 80% and then harder to do anything. But it would probably work for almost any mower base. I find it hard to believe you could make a base significantly cheaper than the mower they found.

And I would argue the math of $35 vs a one time investment isn’t fair. Every machine needs maintenance. The more complicated the machine, the more expensive the maintenance. If your mower runs on ROS, and some package breaks, now you need a software engineer to fix it and they cost a lot more than Jimmy down the street. That all changes if you are into the DIY aspect of it. But don’t kid yourself. It isn’t about saving money. It is about enjoying the making.

1 Like
6

In Germany farmers have to legally have a strip next to every field with flowers for the bees. Pretty neat. It’s called “Blühstreifen”.

6 Likes
7

Yeah, but clover sucks to walk on. I still chase kids/dogs around the yard and I typically run barefoot.

My yard isn’t a putting green by any stretch of the imagination, but I try to at least keep it green.

The field beyond the fence line is left to its own devices. I mow it just to keep a buffer from the corn field for snake and mice reduction, but I don’t worry what’s growing there. The wife and I are thinking about spreading wildflowers over the entire field and just letting it go.

2 Likes
8

Yes, things do break, I agree. Many of the projects I make do break, but since I built them, and programmed the microcontroller boards, if anything does break, I can fix it. I also have a soldering rework station and can replace the SMD chips if needed. In the Navy I troubleshot boards all the time to the component level. So, an investment at first, but it will pay for itself. And yes, I do it for the making. I’m just not very good in CAD.

As for the RTK corrected signal, the antenna would be on my house so I don’t think I’d have to much problem with the signal bouncing. It’s data. So as long as it gets to the OpenMower Bot, it will be accurate enough for a lawn. But thanks for the suggestions and comments. I didn’t think anyone would. Cheers!

1 Like
9

The antenna on the roof measures the disturbance due to things like air pressure or clouds or whatever atmospheric noise there is. Since the base station doesn’t move, if the GPS signal says it is 10cm to the north, it sends a correction of 10cm south to the gps on the robot.

The robot has it’s own antenna and that correction, which should be the same since the atmosphere doesn’t change much between your roof and your yard.

But if the gps on the robot also has trees between it and the satellite, or it is getting a reflection off of your house, those sources of noise can’t be fixed by an rtk system on your roof.

I’m not just guessing here. I am a software engineer who has been making uncrewed ground vehicles since 2006. I have used rtk and ros quite a lot and experienced these problems while sitting in 15 ton robots.

I definitely don’t know everything though. New solutions are found all the time. These days, we mostly just pay $15k for a good nav box that includes a good solution and pay a specialist to fly out to fix it if we have any problems. So I am not up to speed on the latest stuff.

1 Like
10

Hmmm, after your mention that I probably couldn’t make the mower cheaper than the Yardforce Classic 500 used in the OpenMower project, I looked it up. It is available at Amazon DE, Germany for about 358.57 Euros PLUS 213 euros shipping. That’s about $611. and after a brief Google search, I can’t find it here in the US. Any robot mower I can find in Ohio at a Home Depot or Lowes costs about $800. I think I can beat that if I make it myself. Yes, I would definitely use an IMU just to make sure if it was tilted too much from a kid trying to lift it or something the mower blades would stop, as well as help guide it through it’s path. And touch sensors as well. I work in the Parking industry on automated parking equipment. When covid hit, we installed “Touchless” buttons on all of our equipment so you wave your hand in front of the machine to get a parking ticket instead of pushing the button. Those have an adjustable sensitivity and I can get them for $5 each. If it did cost $800 to make the mower, I would know everything about it and be able to get replacement parts cheaper, I think. That’s been my experience anyway. But your point is well made and a must be considered. Thanks for your thoughts.

1 Like
11

Thanks for replying, Jeff. WOW! You’re the person I need to talk to then. I have found that my DJI Mini 2 drone typically can see about 20 satellites around my house. Then Sparkfun Zen-F9? module receives signals from the US, Chinese and Russian satellites and, after sending the files off to the Canadian Geodetic services office to correct it’s location, is said to have 1cm accuracy. I’ve ordered an L1/L2 antenna and should receive it in about two weeks. I’m going to build the Sparkfun base station first and put my site on the RTK2Go service for others to use. My cousin is a surveyor in Kentucky and they use Trimble RTK stations and rovers. He doesn’t know how it all works and I’m just learning. I’ve also been watching a guy on YouTube who is going to UC Davis that uses Drones for agriculture. So, I know hardly anything about RTK and GNSS and the lot of it, but I AM an RF communications technician with over 40 yrs experience. Including LF up to Microwave and satellite comms. So I do know about that side of it. I could put all of the RTK receiver stuff in a small RC car and see if that tracks around the yard well enough. That would prove out the guidance portion of the system with inexpensive BLDC motors. Then, if that works, I could then move onto making a mower. I may ask your advice on some things along the way, Jeff, if that’s OK with you.

2 Likes
12

Sure. Just let me know where to send the bill :slight_smile:.

Joking aside, I will help when I can. But I only do it in public and don’t expect too much on really advanced topics. My professional experience often results in us just buying our way out a problem so we can focus on the problem we are paid to solve. The stuff I am really focused on (path planning) is not something I can share, because it is proprietary at least.

Here is an old video (these demos were 2013ish) of some of my work (I didn’t do it myself, there is a huge team involved with something like this):

(You can safely mute it)

3 Likes
13

Just great, more people giving robots lethal weapons…
Sweeper_Original

1 Like
14

I would love a robo mower, and a smart one would be really nice. Surprisingly though my super cheap dumb robo vac does a great job, might take it some extra time but it bumping around works well. I try to mow my lawn from different directions each time (grandpa told to so I do), so maybe a dumb mower isn’t that bad, unless you have a whole ton grass to cover. It would be cool to have one though.

Maybe I should start by looking into what the open mower project has going on. Can I make it drive between my 3 small grass patches.

2 Likes
15

While moving though? I thought that was only for a stationary sensor.

16

In a wide open field with no nearby buildings, it will be very accurate. Easily 1cm, even while moving. But there are a lot of other sources of error.

1 Like
17
18

How to piss off the FDA…

2 Likes
19

The discussion regarding GPS reminded me of the time when my drone randomly darted toward a crowd of onlookers and crashed hard at their feet… this while in a solid GPS hold hover. It was an APM 2.5 (arducopter 2.6 or so) using an older ublox, but pretty high end at the time. This drone had previously been very reliable, and accurately flew many autonomous missions from takeoff to land. It appeared the incident was caused by a solar flare… this was in the middle of a large plot of farm land, and flare trackers indicated a lot of activity that day. After investigating that incident the devs implemented “GPS safety features”, which were basically sanity checks so the drone won’t suddenly “break right” when a GPS coord changes in an unlikely way. I’m guessing something like that would work well to make GPS glitches less dangerous for a slow moving mower, but I think some optical AI feedback would be good to avoid mowing over large obstacles etc. Then if you go long enough between mowings, AI might also be able to use it’s own lawn track for planning, hehe.

20

DP, sorry… automating mowing with a robot would be a fun tech project, but I think just a remote control mower would be far more reliable and cheaper to implement. Maybe add some FPV for safety/convenience. It would also take most of the labor out of it for those who are not able bodied. One could retrofit servo steering and clutch operation on my larger gas powered mower… add electric start (and power bag empty if you are bagging), and you’ve actually made one of the classic chores way more fun!