Very frustrated youtube took so long to share your channel with me. Love all of your projects and the in depth tutorials. You're awesome Nathan! Keep building!!!!
Now that you have it ready to GO -- might be a good idea to consider some (Safety) ways for it to STOP. You could put bumpers around it so that it cannot run into anything or anyone. At least not very hard. Anything touches the bumpers opens a Normally Closed kill circuit. And maybe some of the big Emergency Stop buttons on top for folks who may have to shut it down in a hurry. Those are not complaints -- just ways to make a good thing better. Thanks for the build, the video and sharing the details.
seems like a good way to introduce reliability issues. how much could the bot realistically damage? that blade is unlikely to make it past the frame, and what sort of folks to you see it encountering in your backyard?
@@xbenng Safety is suppose to reduce reliability to the point it can be externally stopped. It is the INTENT. We sometimes call that Fail-Safe .. which is to say that it Fails in a Safe manner. As opposed to Fail Sorry. Another somewhat more complex method, but considered safer yet is a "Stay Alive" signal. Absent the "Stay Alive" or All Good signal, it goes to an orderly shut-down or Safe mode. An example of this might be a boundary around the zone(s) of operation. Anything crosses into the safety area, and it shuts down.
Wow, you pulled off an obtainable, and actually USABLE automated mower. Excellent job explaining it, providing components, and sources. Thank you, thank you, X3! I'd like to think that after I order the parts, I may actually build one.
Outside of the large area to mow, we have the same goal. I happen to have no skills yet with welding and such. I then opted to use wood for my chassis. Instead of using scooter motors, I'm re-using some e-scooter hub wheels that have the motor in them. I thought about using parts of mower, but figure it'll help me to not use any gas, if I just power another motor with a string like a trimmer to cut the grass. The idea is that eventually I can move that motor up and down to control cut height. I love how "simple" your auto-mow looks, but that it works well. Kudos to you! I'm inspired...again.
Dang, I’ve been toying with the idea of modifying my push mower to be an autonomous robot for over a decade and a half. Glad that someone went and actually did it. This is amazing. I actually have some experience myself with both Ardupilot and the Cube Orange Plus, courtesy of a little challenge put together by NASA that the robotics club at my community college (along with 5 others) participated in. We were tasked with building an autonomous quadcopter. Fun stuff. I do have a couple suggestions. First of which being, consider covering that platform up top with a waterproof shell. That way if you’re suddenly caught out in the rain, you’re not out a thousand dollars in hardware. Especially since the Cube Orange Plus isn’t waterproof, and I’m sure your ESC isn’t either. Second (though this might not be an issue for you personally because it looks like you’re just mowing flat land), I would either reduce the height of that platform, or space the wheels farther apart to give it a wider base. This will help reduce the likelihood of it tipping over if you have to mow on a slope
Good suggestion on the ESC cover! I had some water get on it and it shorted into "on" position, so it just took off on its own. I had to pull power to stop it and dry it off. Good call on wider wheels too for big slopes. It's pretty flat here, but certainly needed if it's used on hills
Yes, I think this page describes how that looks: ardupilot.org/copter/docs/common-gps-for-yaw.html I wanted a single gps/base station correction because I used the same setup on a drone for spraying my fields, so I reused that setup there.
@@nathanbuildsdiyYou should replace tired wheels with apron wheels, which are beneficial to improve your vehicle's off-road function. Especially in lawn surrounding.
The compass is the much more stable and applicable variant, because GNSS has significant noise (wighout RTK) and thus direction calculations will be imprecise.
Thanks for sharing this. This build has inspired me to build my own, fully electric, zero-turn kinematics. Aiming for USD$2k bom, and will be documenting it on GitHub.
Very cool! You should sell this as a kit with all that electronic and computed parts and motors. I think most people can build the chassis and mower themselves, but as soon as it comes to the electronics and controls, it becomes difficult, at least for analogue hobbyists like me.
I get a kick out of building it and making it work, but I'm not so interested in selling it. But I'm hoping other entrepreneurial viewers will do that!
@@nathanbuildsdiy I have an impulse to congratulate you 👏👏👏👏for this giant project but I am wondering and amazed about how one can accumulate tons of this valuable and accurate knowledge and skills, is it about your great university or company with lot of experience? , but as I can guess the your power ,you may be an Mechatronics guy 💪💪💪💪
Dear sir Nathan, thank you so much for your tutorial. Inspired by this video, I made my own automow (but mechanics is different) and it is working via RC remote control. Really grateful to you. Now, I am working hard to set up the RTK GPS and finding it tuff to NTRIP server in southern part of India.
Great work! Is there much agriculture in your area? Do farmers use large machines? If so I bet you can talk to them to find out what base station network they use. Most require a username/password, which is usually free to obtain. The machine manufacturer charges to use RTK, but the actual base stations are free to anyone. That's how it is in my area. Farmers use LEICA network base stations on their tractors and combines and I got a username/password to use that network. If there is no network available, then you could create your own, but it requires this: irlock.com/products/here-base-m8p And here's some documentation on setting up the base: ardupilot.org/copter/docs/common-here-plus-gps.html#base-module-setting-using-mission-planner
Sir @@nathanbuildsdiy Here in India, precision agriculture is not widely available yet (especially in my locality). So I will check setting up my network with help of your links and keep you posted
Great project mate. I am also working on a similar kind of thing but due to some budget limitations i am trying to make a mower using Arduino UNO that remembers the path its being trained on using an android app. As a mechanical engineering student i must say the its a very impressive and sturdy frame design. Keep up the good work. Hope to see more projects like this.👍👍
Arduino UNO is a great platform, raspberry pi zero 2 may cost even less and have a bit more horsepower. I use the UNO in this build to control a bunch of motors, code for that is linked in the description: th-cam.com/video/Q7A3PlqpgYs/w-d-xo.html The Cytron motor controllers are really simple to use with the UNO I've found that depending on the condition of the ground (muddy, long grass etc.) the time that it takes to turn 90 or 180 changes, so putting a compass on your UNO may help you figure out if you've turned the appropriate amount.
you will really need a gps correctionfactor, because of the track changes, like tops and things that the wheel stuck , and it changes the course, even if it changes like tiny degree, it will change the position in long distances, and it will cause portions of not cuted grass. Sorry about the english.
Thanks for your great video and especially the circuit diagram sketches, I built a remote-controlled petrol lawn mower last winter (Igor 3.1 ^^), which is basically for the really rough stuff. Next winter I also want something autonomous for fine cutting.
I would try to make a small riding mower engine bolt to that mower deck, not for the extra power but because riding mower engines have a dynamo under the flywheel that could help charge that 12v battery on the go, sweet build 👍
Probably the power of the dynamo wouldn’t be enough for powering the electric motors and charging the battery IMO .Better to think about fitting some relatively small car alternator as from Geo metro / Suzuki swift for example . Also if you are good with electronics, you can make controller which turns the alternator in to brushless motor just for starting the lawnmower IC motor so it will serve double purposes . Also you have done great job overall but the mechanical part isn’t very good . So IMO try on the next iteration Brushless wheel -motors from hoverboards . They are 250-300W each ( direct drive) and the speed controllers from the hoverboard can be reprogrammed and reused for your application as well( there are info on the net how to do it ) . Mechanicaly will be much simple and reliable this way and most probably cheaper. You can buy second hand hoverboard dirty cheap and you will get 2 brushless motor-wheels and 2 speed controllers from it . The way ( and the details)how they are mounted to the hoverboard will be also useful for your application.
Can't find GPS precision information on your Cube component, but I have a question for you. What accuracy/precision do you get from your GPS signal? Normally these signals are in the 2 meter precision range, which is hopeless for this mowing application. But there are realtime GPS correction streams available depending on your government and region which can get you down to sub 2cm precision at around 1 solution per second which is ideal for mowing precisely. Specifically: Real Time Kinematics. RTK is short for real time kinematics. A GPS receiver capable of RTK takes in the normal signals from the Global Navigation Satellite Systems along with a correction stream to achieve 1cm positional accuracy. This is again, dependent on governmental infrastructure and whether the signal is available to anyone or requires a license presumably and is encrypted. Edited: I see you go thru the RTK/NTRIP settings for GPS correction stream data at 17:30. That answers my questions. Thanks for including that detail. Most important.
RTK GPS using base station correction (same one the farmers use here for their tractors etc.). That's how I get accurate enough. Ardupilot makes it easy!
@@nathanbuildsdiy Yeah, it's really cool what you can achieve nowadays with GPS. No need to spend megabucks on these overpriced gps mower contraptions, when you can build your own for a fraction of that.
Very cool ideas, not just like the whole thing but stopping the chains from derailing and making it so that it will actually do work for an hour without screwing up
It actually ran for a long time with no chain problems, and then when the grass got long and branches started falling suddenly I had problems so I added the chain keepers long after it was being used
Cool build. I really like the large wheels. I would prefer electric motors on the axle instead of chain drive. The chains appear as if they might catch twigs and branches since it is so open and high. It is an interesting choice to use an ICE for the blade drive. I definitely understand the reasoning. It would have been fun to see an electric cut.
The chains do pick up sticks. In the spring all was well, but in the fall more sticks were dropping and the chains started to derail. I had to add chain keepers (my name for the big circles of metal added inside the wheel) to keep the chains on and that's been good for now.
I love this transmission as its so cheap and most energy is being used from the gasoline anyway. Internal geared hub drives are really expensive or really weak
This is pretty much what I had in my head for mowing my own section, including the RTK GPS. But I never knew about mission planner or ardupilot etc. But since I don't have a house on it yet, that one is on the back burner lol. Sounds like most of the firmware/software is 100%, so it's just implementing the design to your needs.
Crude AF but it proves the concept. Always fun to see 1st or 2nd iterations and then the possibility of refinements. Easy to do from here ;-) Thanks for sharing.
Thanks for watching. It is crude, next iteration will keep the electronics protected and keep debris out of the chain drive. That said, it's been running for a long while now with only minor repairs... it's a little tank.
@@nathanbuildsdiy I asked my daughter if she would be interested in help building it. She said ok but maybe not exactly like that.. I said why not? She said, well it looks ... and she couldn't think of a word so I said ghetto? and she said yeah, it needs to look better.. I told her to draw up something. :) I was already planning on doing something close to it.. I hadn't decided on the wheels yet and was planning on 2wd, but you probably need 4wd.
Sweet project utility>>>>looks, approved you remind me a lot of myself, if I was more handsome, less lazy, had a welder that functioned, and time to do sweet projects. Subscribed, thanks for sharing
twisted pair works much better with differential drive, which eliminates common mode signals. The twisted pair will cancel out a fair amount of induced common made, but any residual is still fully presented to a single ended input.
You should move the accelerometer down to the axle height and add some gyros. This way you can ignore GPS if the fix isn't great and do dead reckoning until GPS is good again. Should be good enough for what you're doing :)
I got a mower with fwd wheels there is a pulley with a belt on a main shaft ,you could have add a pulley there or use mower like that and connect an alternator there
sweet build i got one suggestion get a 220 welder i bought a el cheapo off amazon and do not regret i runs fin off a old school generator no need for inverter type to much hype on them in my option oh and said welder will run off 120 and with .30 wire it will burn in good and i was stick welding with AC 22-volt tombstone welder fer years. get gas tanks for the mig its sweet stuff no looking back after that .should have did it years ago
Not bad for a proof of concept. I like the chain guides on the rear. One thing though, I think the audio got messed up while editing the video. After 10:12 I can only hear you in my left side of my headset right around the Schematics portion.
You could add that and get somewhat longer run times, but it chews through power faster than a solar panel could recharge it, so it'd only add minimal total run time before a recharge
Yes, solar this small sadly gets you nowhere. A small generator attached to the engine would. I think it would still need some routine to catch up the charge because of that small engine has not enouch power available to run continously. Next thing is You don't want to mow if the sun ist suitable for solar. So that would need stationary big solar and fully unattended mow and return to charge capabilities.
Yellow or any color really should work fine. This thing is slow and doesn't take hard corners or anything like that. A fast flying drone might care about which cube is in there, but this one doesn't!
I am impressed, so much so that I build my own rover to do robotic farming. All the GPS stuff was based off your partlist and block diagram. However I was not able to get my rover to reverse. The ESC type is "BrushedWithRelay". In the Motor Test, I was able to spin both L and R wheels with motors test C and D. However, when I set "Throttle %" to a negative percentage, it would not output the relay pins on Aux5 and Aux6 to go High. I was able to get GPS RTK float, and Auto drive the rover forward, but not backwards. The only way to get reverse, was to externally drive +3.3V to my transistor that engages the coil, which flips relays for the motor +/- connections. The Ardupilot Aux5 and Aux6 should do that, right? Am I missing something in Full Parameters List (like Relay_Pin and Relay_Pin2 = 54 and 55)? Thanks!
Dude!! This is great stuff!! Much appreciate you taking the time to document, and share your knowledge with everyone. Quick question if you have time. Would it be possible to drive my rc mower manually, from my RC controller, with the cube passive but online, then add my waypoints in the spreadsheet as I progress? Theoretically, one time cutting manually using the RC transmitter i'd have it mapped, then turn my auto/Man switch to Auto, switching from RC mode to Autopilot, and operate off the GPS. Oh yea you mentioned distance in the ethernet cable being a consideration...Do you know the max length by chance? Id need to run at least 5' from the rear up to the controller in the front. Can't thank you enough for your hard work, and allowing everyone to benefit from it! Kudos! Thanks Dave
Certainly - you can either download logs from your run and parse through those or (probably easier) just pull up the "data" tab in mission planner while the mower drives around. The current GPS coordinates are at the bottom left side of the map view in that tab, just note them/save them off in a file somewhere and use them to generate your map file.
Have you thought about using two sprockets for each wheel and two bike chains on each side to make tank treads? It may help with off-road traction and any unexpected objects like sticks.
I certainly did consider... I actually wanted to 3d print some treads. I bet bike chains on the ground would pick up sticks etc, but it would be worth a try. If you do try, let me know!
Maybe you mentioned it and i missed it, but does the program have any type of safety stop mechanism, say if the mower has some mechanical failure and wanders off the gps track? Is there a way to program it to stop and/ or shut down? I live near a highway and would want some sort of safety feature that would prevent it from roaming out on the road.
This is a good idea - I talked about it in another comment somewhere in here. Basically I think it should be added - it'd be fairly simple. The cube orange has a arm/disarm switch that plugs into the usb port. That could be wired to "bumpers" on the outside edge of the mower so when they're hit it would disarm and stop moving
amazing job with this setup! some water proofing could make the electronics last longer, maybe a 3d printed enclosure or anything else which protects from water.
Absolutely. It's on the to do list. I actually had it running last summer when a storm blew in. A drop of water hit the motor control in just the right place and suddenly the right motor was stuck "on". Luckily I was already on the way to turn it off when it started driving off into my field!
You might throw a solar panel and a charge controller on top to help that battery not drain too low too many times. Or better yet, see if you can add an alternator onto the engine.
Does the mower download the RTK updates in real time or are the RTK updates downloaded to your laptop and then transferred to the mower? Great description!
It's near real time, but there is some latency. The laptop does all the communicating with the RTK network because that's done over the internet. So the drone gets position info, sends that over the 900MHZ radio to the laptop, the laptop talks to the base station over the internet and relays corrections back to the drone over 900MHz. That's all handled in mission planner though, so once you're hooked up to a base station you're good, no need to mess with the details.
Hi Nathen, Thanks for the awesome project , Right now i am collecting the material to build one. I have a question? Since Motors are 24volt rated why are you using a single battery of 12volt to supply the voltage to dc motors? How is it going to support the 24volt dc motors on 12volt battery?
I originally was going to use 2 12V DC batteries in series for 24V, but when I ran it I found it was plenty fast with 12V and I could charge the second battery while the first one ran, which I liked even more. That's why I ended up using 12V.
Nathan how soon are you coming out with your updated mower. I’m very interested in building one for my grandmother before she passes. She constantly pays people to do her lawn. I’m not sure if you have any contact information for help with my build? I was also wondering do you have to manually start this mower every time it starts? Could you use an electric mower and connect it to the car battery?
This is an incredible project. I made one, but mine is remote controlled. Is it possible to use a sabertooth motor controller, or do I need the controller you used? I don't know very much about autonomous. My setup is ready, all I need is the autonomous parts. Thank you for sharing your talent.
I just did a search on sabertooth pinout and it looks like there are 2 main and 2 aux pins on most of them (dual motor controllers anyway). Do you know if the main is a PWM signal (amplitude basically) and the aux is direction? If so it should work with the cube orange/ardupilot just fine.
@@nathanbuildsdiy Thank you for replying. I have to look into the pwm and which output is which. I use the sabertooth with a remote control. However, I want to venture out and learn more by converting it as autonomous. Again, thank you.
Other than CG issues, is there enough traction and torque to climb a 15° hill? Perhaps with more aggressive tires? With your setup, is it effectively 4WD?
Totally 4WD and these motors are very beefy. I had it climb up a tree and flip on its butt before I got the maps correct, so I think if you had enough tread on your tires a 15 degree slope would be fine. Perhaps put weights on the outside hubs (like we do on riding mowers) to prevent rollovers when making turnarounds
Thanks. I am investigating more aggressive wheels for the hills, your amazon link for the wheels go to a 26" wheel. The ones you used look smaller, is that the case? @@nathanbuildsdiy
Yes, certainly. Pixhawks (closed source and less quick processors) can be had in kits on amazon. Their GPS's aren't as good as here3 though. I *think* it's possible to get a here3 talking to one of those, but didn't try it myself
Did you make the frame so tall to get the electronics away from the spark plug and coil; or is that an issue? I would like to make a more compact version.
It does isolate the electronics a bit, but I think the electric motors were a bigger issue. I made it that tall because I wasn't sure how the whole thing would turn out when I started :) I'm betting you could cut down the height and be just fine as long as the gps/compass are far from the electric motors, and as long as you're able to get your mower into the frame (I have to turn mine diagonally to insert it from the back).
Any update on your next generation mower? I am trying to build an autonomous riding lawn tractor which already has alternator but has Ackerman steering instead of skid steering as in your project. Used riding lawn tractors can be bought for a couple hundred dollars. I would love to use your setup with Here3- Ardupilot rather than AGOpenGPS
Hi! Sorry for the delay in response, been busy with family over the holidays. I'm currently working on a version 2 of the weeder and drone sprayers that I built and haven't updated this one yet, but it's my goal to get it this summer. Version 1 is still plugging along, though I had to replace the car batteries powering it Hit subscribe if you haven't already, then that build will pop up in your feed once I (finally) get to it.
Nice work! I live on 12 acres. This could be really helpful. Have you considered a battery charger (alternator) to keep the batteries full all the time? Thanks for the explanation. Glad I stumbled on this video.
Pour entraînement pas de problème car en mettant que sur deux roue sa suffit moi c'est le guidage gps qui m'intéresse le reste c'est de la bricole se qui compte c'est le guidaje droit avec changement de voie et arrêt dans la zone déterminer toujour votre vidéo est très instructif
Wonderful stuff, man! Didn't know the open software was so good now. Even training the AI for your freznel weed killer seems to be approaching dummyproof. The only upgrades I would make are: Weld on a sorta cattle-guard to push away the big rocks and branches .. Larger tires just for traction, because i'm building a chariot behind mine to pull me around the property like a retired drunken Spartacus. Give your map to a drone to do a quick scouting ahead.. Honestly forget even the mower.. if one could just come up with a pathfinding diy platform rugged enough to carry a couple of chainsaws and a gas can/generator water/lunch/guns .. There's nothing worse than losing an hour+ of daylight just because you forgot/need a pair of needle nose. It's gonna break, of course, like everything on farm will.. I'd much rather have to deal with a broken or jammed up chain than dealing with anything like tank treads unless the property absolutely requires it.
I love the project! A suggestion would be that during the voiceover portions (like at 4:33), your voice only comes out of my left headphone and it's a little bit annoying
You're absolutely right. I didn't realize it was mono audio on the voiceover until after I had published. TH-cam won't let me upload a new video without deleting this one. All future vids will be stereo audio.
I'm doing my build now and looked at the video and schematic again closely. For Aux out 5 & 6, what did you configure in Ardupilot for this setup (wheel direction) ? I'm using the bigger Cytron MDDS30 and it is slightly different. Also what did you set for MOT_PWM_TYPE? I'm using wheel chair motors for my setup.
Please let me know were and how to download the software for gps controller. It is Wonderful Assembling and you are genius. Please let me know if the gps section we can do it cheaper from china cart gps. Thanks, I love it.
Thanks for the comments! Here's where to get mission planner, the software for the base station (it's installed on your laptop): ardupilot.org/planner/docs/mission-planner-installation.html Here's where to get ardurover firmware - it's installed on the cube orange (or pixhawk) autopilot: ardupilot.org/rover/docs/common-downloads_firmware.html There's good documentation about getting these set up here: ardupilot.org/rover/docs/apmrover-setup.html
LOL, well played. I didn't realize the audio was mono when I uploaded and I can't upload the stereo version without deleting this video! I'll fix for future videos
They run at a lower speed basically. I tried running at 24V initially but the whole thing moved too fast for the mower to keep up. When I tried at 12v it was happy.
That's great! Let me know how the build goes, if you've got any questions post a new comment (I always see the new comments, sometimes miss replies). We can connect outside TH-cam for any help/questions. As for 24V - yes, you totally can set the max speed in mission planner (it's on the basic tuning page to the right of the P I D settings I think). When I started 24V testing I didn't realize that was there. I tested 12V and it was a speed I liked so I kept it and got into the habit of charging 1 battery while the other operates, so I've always got one ready to roll. Much later I realized I could have stuck with 24V and set a max speed but I had already developed a routine. If you're going to build here's my list of ideas for the next version, pulled from the comments here, if you want to try any of them (no need, but I'm thinking I'll add next): - add a front bumper to stop if we hit something (connect to the arm/disarm button that goes to the USB on cube orange) - Add a small alternator on mower to rob a bit of power and charge up battery - Add a chain tensioner. - Put a pin in the axle so it can be removed rather than welding that inside washer to the axle - 2 gps for yaw/no need for compass - But man... this is more $ and I just can't pull the trigger
I built a robot similar to this a while ago and found that the GPS resolution was not fine enough. Non military GPS can have a variance of up to 3 feet. How did you prevent the mower from running over Grandmas flower bed?
GPS RTK - it's base station corrections which get you to sub centimeter accuracy. I discuss how to set that up in misison planner with the here3+ GPS in the middle of the video. That's also the reason for the more expensive GPS unit, it can do RTK.
I planned to use 24v (2 12 in series) but realized it was a good speed with 12V and I could charge the other battery while one was running. Now I'm working on an alternator/regulator to recharge the battery while it runs and that's a 12V system so I'll stick with that.
Yeah, I thought about doing that and couldn't find any used ones for cheap. You want to have rpm's fairly low - these motors do the gearing down inside the enging housing - something like 100-200RPM depending on the size of tires you have, and plenty of torque at the low speeds.
To get Aux 5 & 6 to work for directional control, did you have to set any parameters? I have the same setup as you. Pin outs on the controller and cube have been double checked and the cabling is good. However I can not get the motors to change direction. Any suggestions?
Sorry for the delay! Been busy with kids starting school. Lets check the settings - Initial setup tab -> Mandatory hardware -> servo output -> throttle left/right are channels 1/3, correct? Config tab -> basic -> Brushed with Relay You're connecting to the AUX 5/6, not the main out 5/6, correct? Next go to the Config tab, under advanced (I think?) and search on the word "aux", it should pop up what those are set to. Next up... I suppose check with a multimeter connected to the output pins to see if aux 5/6 are actually changing voltage.
@@nathanbuildsdiy thanks for the reply. I will run through your suggestions to see what happens. If I select ‘brushed BiPolar’ then I can get the mower to turn. If I select brushed with relay then when I command a turn both wheels still go forward. I do have it set up for rc control also, that might be a factor also. Anyway, I was just looking for a quick fix, not product support. I have the winter to get the bugs out. If you would like here is a link to my video of the mower, I referenced your channel in the video and included a link in the description. th-cam.com/video/mMNRO0pRbzY/w-d-xo.htmlsi=ES4XU3riRxZxvhAf
Kudos for this build. I'm impressed how this works. But I would never run a live mower withoud additional sensors, that avoid objects PLUS constant observation. So I wouldn't benefit from this build 😞
Totally understand. I'm fortunate that the farm is quite isolated and surrounded by trees and fence. I'm the only person or pet around when its running and I work nearby so i can keep an eye on it.
Would love to see you put a solar panel on that. Electric start mower. Battery dies it cuts mower charges to a specified point cranks up the mower and keeps going. Would be the next step to a totally solar autonomous mower just need to go lithium and electric mower motor. It would be a mechanical goat.
I like the idea, but I bet the size of the solar panel needed to meet the power demands would be pretty large. I think next build I'm going to add a small alternator to the mower to power the battery
Yeah, all that is handled in ardupilot. I set waypoint "size" as small as possible, so it doesn't try to make a big U turn but instead pivots in place.
2 kewl. I have built a few RC mowers and would like one to be autonomous. Does this require any kind of ongoing subscription fee or is it a one time cost?
No subscription cost to connect to the base stations here in Iowa. Farmers use them all the time for their combines/tractors so it may be different in your area, but all systems I've seen in surrounding states are also free
Yeah, they are large, but were a good price on Amazon so I went with them, plus the frame and car battery are pretty heavy, so the extra power isn't for naught
It's a barn I built myself. The posts and rafters were all cut from trees that I removed from the surrounding fields (they're a type of weed tree around here, always going where they shouldn't). The barn boards were salvaged from my grandma's barn that fell in a big storm about 15 years ago. They're about 140 years old! I was actually thinking of a video on how to build that, I'll make it now!
I like the idea, but I bet the size of the solar panel needed to meet the power demands would be pretty large. I think next build I'm going to add a small alternator to the mower to power the battery
Very frustrated youtube took so long to share your channel with me. Love all of your projects and the in depth tutorials. You're awesome Nathan! Keep building!!!!
Now that you have it ready to GO -- might be a good idea to consider some (Safety) ways for it to STOP. You could put bumpers around it so that it cannot run into anything or anyone. At least not very hard. Anything touches the bumpers opens a Normally Closed kill circuit. And maybe some of the big Emergency Stop buttons on top for folks who may have to shut it down in a hurry. Those are not complaints -- just ways to make a good thing better. Thanks for the build, the video and sharing the details.
All good points, I think next version will include that. The bumpers + kill switch can all be one circuit that connects to the cube orange switch.
another safety could be when tilted beyond a certain angle
seems like a good way to introduce reliability issues. how much could the bot realistically damage? that blade is unlikely to make it past the frame, and what sort of folks to you see it encountering in your backyard?
@@xbenng Safety is suppose to reduce reliability to the point it can be externally stopped. It is the INTENT. We sometimes call that Fail-Safe .. which is to say that it Fails in a Safe manner. As opposed to Fail Sorry. Another somewhat more complex method, but considered safer yet is a "Stay Alive" signal. Absent the "Stay Alive" or All Good signal, it goes to an orderly shut-down or Safe mode. An example of this might be a boundary around the zone(s) of operation. Anything crosses into the safety area, and it shuts down.
@@philtimmons722safety is boring world has enough humans and could lose a few with out notice.
Wow, you pulled off an obtainable, and actually USABLE automated mower.
Excellent job explaining it, providing components, and sources. Thank you, thank you, X3!
I'd like to think that after I order the parts, I may actually build one.
It's a fun project, and the ardupilot platform can be the brains for just about any robot you can dream up
Outside of the large area to mow, we have the same goal. I happen to have no skills yet with welding and such. I then opted to use wood for my chassis. Instead of using scooter motors, I'm re-using some e-scooter hub wheels that have the motor in them. I thought about using parts of mower, but figure it'll help me to not use any gas, if I just power another motor with a string like a trimmer to cut the grass. The idea is that eventually I can move that motor up and down to control cut height. I love how "simple" your auto-mow looks, but that it works well. Kudos to you! I'm inspired...again.
Thanks for the walk-through on the GPS and piloting software, that I haven't gotten to jump into yet.
Good luck! let me know how it goes
Dang, I’ve been toying with the idea of modifying my push mower to be an autonomous robot for over a decade and a half. Glad that someone went and actually did it. This is amazing.
I actually have some experience myself with both Ardupilot and the Cube Orange Plus, courtesy of a little challenge put together by NASA that the robotics club at my community college (along with 5 others) participated in. We were tasked with building an autonomous quadcopter. Fun stuff.
I do have a couple suggestions. First of which being, consider covering that platform up top with a waterproof shell. That way if you’re suddenly caught out in the rain, you’re not out a thousand dollars in hardware. Especially since the Cube Orange Plus isn’t waterproof, and I’m sure your ESC isn’t either.
Second (though this might not be an issue for you personally because it looks like you’re just mowing flat land), I would either reduce the height of that platform, or space the wheels farther apart to give it a wider base. This will help reduce the likelihood of it tipping over if you have to mow on a slope
Good suggestion on the ESC cover! I had some water get on it and it shorted into "on" position, so it just took off on its own. I had to pull power to stop it and dry it off.
Good call on wider wheels too for big slopes. It's pretty flat here, but certainly needed if it's used on hills
With a vehicle frame that large, you could use two GPS for yaw and forget about the compass hassles
Yes, I think this page describes how that looks: ardupilot.org/copter/docs/common-gps-for-yaw.html
I wanted a single gps/base station correction because I used the same setup on a drone for spraying my fields, so I reused that setup there.
Wait, that’s the reason large autonomous vehicles have two GPS?? 😳
@@nathanbuildsdiyYou should replace tired wheels with apron wheels, which are beneficial to improve your vehicle's off-road function. Especially in lawn surrounding.
@@djiang2738 what is an apron wheel? Google didn't help
The compass is the much more stable and applicable variant, because GNSS has significant noise (wighout RTK) and thus direction calculations will be imprecise.
Thanks for sharing this. This build has inspired me to build my own, fully electric, zero-turn kinematics. Aiming for USD$2k bom, and will be documenting it on GitHub.
Very cool! Make a video about it, I'll watch!
This is awesome. Heck no one would fault you for having a robotic simply because the word “farm” is in the same sentence as robomow.
Thanks! I'm always looking for ways to make farm work more efficient. It's like my kids - always takes all available hours in my day :)
Very cool! You should sell this as a kit with all that electronic and computed parts and motors. I think most people can build the chassis and mower themselves, but as soon as it comes to the electronics and controls, it becomes difficult, at least for analogue hobbyists like me.
I get a kick out of building it and making it work, but I'm not so interested in selling it. But I'm hoping other entrepreneurial viewers will do that!
@@nathanbuildsdiy I have an impulse to congratulate you 👏👏👏👏for this giant project but I am wondering and amazed about how one can accumulate tons of this valuable and accurate knowledge and skills, is it about your great university or company with lot of experience? , but as I can guess the your power ,you may be an Mechatronics guy 💪💪💪💪
Dear sir Nathan, thank you so much for your tutorial. Inspired by this video, I made my own automow (but mechanics is different) and it is working via RC remote control. Really grateful to you. Now, I am working hard to set up the RTK GPS and finding it tuff to NTRIP server in southern part of India.
Great work! Is there much agriculture in your area? Do farmers use large machines? If so I bet you can talk to them to find out what base station network they use. Most require a username/password, which is usually free to obtain. The machine manufacturer charges to use RTK, but the actual base stations are free to anyone. That's how it is in my area. Farmers use LEICA network base stations on their tractors and combines and I got a username/password to use that network.
If there is no network available, then you could create your own, but it requires this:
irlock.com/products/here-base-m8p
And here's some documentation on setting up the base:
ardupilot.org/copter/docs/common-here-plus-gps.html#base-module-setting-using-mission-planner
Sir @@nathanbuildsdiy Here in India, precision agriculture is not widely available yet (especially in my locality).
So I will check setting up my network with help of your links and keep you posted
I like RC but this is on another level of automation. Great development work.
Thank you very much!
Great project mate. I am also working on a similar kind of thing but due to some budget limitations i am trying to make a mower using Arduino UNO that remembers the path its being trained on using an android app. As a mechanical engineering student i must say the its a very impressive and sturdy frame design. Keep up the good work. Hope to see more projects like this.👍👍
Arduino UNO is a great platform, raspberry pi zero 2 may cost even less and have a bit more horsepower.
I use the UNO in this build to control a bunch of motors, code for that is linked in the description: th-cam.com/video/Q7A3PlqpgYs/w-d-xo.html
The Cytron motor controllers are really simple to use with the UNO
I've found that depending on the condition of the ground (muddy, long grass etc.) the time that it takes to turn 90 or 180 changes, so putting a compass on your UNO may help you figure out if you've turned the appropriate amount.
you will really need a gps correctionfactor, because of the track changes, like tops and things that the wheel stuck , and it changes the course, even if it changes like tiny degree, it will change the position in long distances, and it will cause portions of not cuted grass. Sorry about the english.
Thanks for your great video and especially the circuit diagram sketches, I built a remote-controlled petrol lawn mower last winter (Igor 3.1 ^^), which is basically for the really rough stuff. Next winter I also want something autonomous for fine cutting.
Outstanding! Hello from Canada. I've been thinking of a project like this for years. The information you've provided is invaluable. Thanks!
Glad it was helpful!
add RC controller so you can drive it around and record the exact GPS track coordinates into the spreadsheet.
Impressive and useful. Thanks for taking the time to do this.
I would try to make a small riding mower engine bolt to that mower deck, not for the extra power but because riding mower engines have a dynamo under the flywheel that could help charge that 12v battery on the go, sweet build 👍
That's an excellent idea. I'm going to add it in a future revision
Probably the power of the dynamo wouldn’t be enough for powering the electric motors and charging the battery IMO .Better to think about fitting some relatively small car alternator as from Geo metro / Suzuki swift for example . Also if you are good with electronics, you can make controller which turns the alternator in to brushless motor just for starting the lawnmower IC motor so it will serve double purposes . Also you have done great job overall but the mechanical part isn’t very good . So IMO try on the next iteration Brushless wheel -motors from hoverboards . They are 250-300W each ( direct drive) and the speed controllers from the hoverboard can be reprogrammed and reused for your application as well( there are info on the net how to do it ) . Mechanicaly will be much simple and reliable this way and most probably cheaper. You can buy second hand hoverboard dirty cheap and you will get 2 brushless motor-wheels and 2 speed controllers from it . The way ( and the details)how they are mounted to the hoverboard will be also useful for your application.
It would be interesting to try this concept out with a riding motor as a complete unit. It’s already got everything but remote control.
Can't find GPS precision information on your Cube component, but I have a question for you. What accuracy/precision do you get from your GPS signal? Normally these signals are in the 2 meter precision range, which is hopeless for this mowing application. But there are realtime GPS correction streams available depending on your government and region which can get you down to sub 2cm precision at around 1 solution per second which is ideal for mowing precisely. Specifically: Real Time Kinematics. RTK is short for real time kinematics. A GPS receiver capable of RTK takes in the normal signals from the Global Navigation Satellite Systems along with a correction stream to achieve 1cm positional accuracy. This is again, dependent on governmental infrastructure and whether the signal is available to anyone or requires a license presumably and is encrypted.
Edited: I see you go thru the RTK/NTRIP settings for GPS correction stream data at 17:30. That answers my questions. Thanks for including that detail. Most important.
RTK GPS using base station correction (same one the farmers use here for their tractors etc.). That's how I get accurate enough. Ardupilot makes it easy!
@@nathanbuildsdiy Yeah, it's really cool what you can achieve nowadays with GPS. No need to spend megabucks on these overpriced gps mower contraptions, when you can build your own for a fraction of that.
Very cool ideas, not just like the whole thing but stopping the chains from derailing and making it so that it will actually do work for an hour without screwing up
It actually ran for a long time with no chain problems, and then when the grass got long and branches started falling suddenly I had problems so I added the chain keepers long after it was being used
man I wanted to build Gps Lawnmower with Inav few years back but I am always busy. well done!
It does take some time to get it going, but ardupilot is a great platform, pretty simple once you get the hange of it.
Nothing short of brilliant. I learned so much in one place. Thanks!
Cool build. I really like the large wheels.
I would prefer electric motors on the axle instead of chain drive.
The chains appear as if they might catch twigs and branches since it is so open and high.
It is an interesting choice to use an ICE for the blade drive. I definitely understand the reasoning.
It would have been fun to see an electric cut.
The chains do pick up sticks. In the spring all was well, but in the fall more sticks were dropping and the chains started to derail. I had to add chain keepers (my name for the big circles of metal added inside the wheel) to keep the chains on and that's been good for now.
I love this transmission as its so cheap and most energy is being used from the gasoline anyway. Internal geared hub drives are really expensive or really weak
Very cool project!!! I hope to see a solar powered one project like this!!! 👍👍👍
This is pretty much what I had in my head for mowing my own section, including the RTK GPS. But I never knew about mission planner or ardupilot etc. But since I don't have a house on it yet, that one is on the back burner lol. Sounds like most of the firmware/software is 100%, so it's just implementing the design to your needs.
Crude AF but it proves the concept. Always fun to see 1st or 2nd iterations and then the possibility of refinements. Easy to do from here ;-) Thanks for sharing.
Thanks for watching. It is crude, next iteration will keep the electronics protected and keep debris out of the chain drive. That said, it's been running for a long while now with only minor repairs... it's a little tank.
@@nathanbuildsdiy I asked my daughter if she would be interested in help building it. She said ok but maybe not exactly like that.. I said why not? She said, well it looks ... and she couldn't think of a word so I said ghetto? and she said yeah, it needs to look better.. I told her to draw up something. :)
I was already planning on doing something close to it.. I hadn't decided on the wheels yet and was planning on 2wd, but you probably need 4wd.
Sweet project
utility>>>>looks, approved
you remind me a lot of myself, if I was more handsome, less lazy, had a welder that functioned, and time to do sweet projects.
Subscribed, thanks for sharing
Thanks for your kind words, I'm laughing, pretty sure my wife would tell me I look scruffy in my work boots :)
twisted pair works much better with differential drive, which eliminates common mode signals. The twisted pair will cancel out a fair amount of induced common made, but any residual is still fully presented to a single ended input.
You should move the accelerometer down to the axle height and add some gyros. This way you can ignore GPS if the fix isn't great and do dead reckoning until GPS is good again. Should be good enough for what you're doing :)
I also ran across this - 2 GPS setup: th-cam.com/video/NjaIKyrInpg/w-d-xo.html
Is that similar to what you're suggesting?
I still have ardupilot 2 APM, thought would be vaporware after 3d robotics went belly up. good to see that community is still going strong!
Yeah, they've had to start supporting a lot of different vendors to keep things going
absolutely fantastic build, lovely setup in its simplicity and function
Thanks!
I’m French and its a bit difficult to understand evrything but it was very helpful and complet.
I got a mower with fwd wheels there is a pulley with a belt on a main shaft ,you could have add a pulley there or use mower like that and connect an alternator there
Is it a push mower? I've got a front wheel drive push mower too, but it can't steer, just pull foward
@@nathanbuildsdiy yes , its a push mower and it can't steer , there is a V shaped belt like on old cars engines
sweet build i got one suggestion get a 220 welder i bought a el cheapo off amazon and do not regret i runs fin off a old school generator no need for inverter type to much hype on them in my option oh and said welder will run off 120 and with .30 wire it will burn in good and i was stick welding with AC 22-volt tombstone welder fer years. get gas tanks for the mig its sweet stuff no looking back after that .should have did it years ago
i love hearing this guy talk in my left ear in the beginning
Not bad for a proof of concept. I like the chain guides on the rear. One thing though, I think the audio got messed up while editing the video. After 10:12 I can only hear you in my left side of my headset right around the Schematics portion.
Sorry about that, you're totally right. I didn't realize the audio was mono, new videos will all be stereo.
Add a solar pannel ?? And would try to add the voltage created from the mower to charge the batterie ??
You could add that and get somewhat longer run times, but it chews through power faster than a solar panel could recharge it, so it'd only add minimal total run time before a recharge
Yes, solar this small sadly gets you nowhere. A small generator attached to the engine would. I think it would still need some routine to catch up the charge because of that small engine has not enouch power available to run continously.
Next thing is You don't want to mow if the sun ist suitable for solar. So that would need stationary big solar and fully unattended mow and return to charge capabilities.
can a Yellow cube do the same job or i necessary need an orange cube? Thanks great project.
Yellow or any color really should work fine. This thing is slow and doesn't take hard corners or anything like that. A fast flying drone might care about which cube is in there, but this one doesn't!
@@nathanbuildsdiy thanks, good tip
awesome video love the complete software overview!
Glad you liked it!
Super awesome. I’ll definitely have to check out what else you have posted.
Thanks for the content
Check the autonomous drone and landing pad, they're similar and pretty fun
@@nathanbuildsdiy on it like a fat kid on cake.
You could also outfit with a mosquito fogger for insect control. Something organic if you prefer.
Ha! Love it
Have it come home and connect to a charger when it runs low on charge. I think bumper rings would work.
Excellent idea, It'd keep it out of the rain too (I put it in the garage now). I'll try for the next iteration
I am impressed, so much so that I build my own rover to do robotic farming. All the GPS stuff was based off your partlist and block diagram. However I was not able to get my rover to reverse. The ESC type is "BrushedWithRelay". In the Motor Test, I was able to spin both L and R wheels with motors test C and D. However, when I set "Throttle %" to a negative percentage, it would not output the relay pins on Aux5 and Aux6 to go High. I was able to get GPS RTK float, and Auto drive the rover forward, but not backwards. The only way to get reverse, was to externally drive +3.3V to my transistor that engages the coil, which flips relays for the motor +/- connections. The Ardupilot Aux5 and Aux6 should do that, right? Am I missing something in Full Parameters List (like Relay_Pin and Relay_Pin2 = 54 and 55)? Thanks!
Did you figure this problem out? I have the same issue. Ended up changing servo_14function to - 1 gpio
This is awesome!
I wonder if you could get rid of the magnetic radiation from the motors with a faraday cloth? Or atleast minimize it...
That'd probably work too, or 2 gps - I think ardupilot supports that and it can handle the EM flux's too
Awesome project. Thanks for sharing all the details
You're welcome! I've got a few more interesting ones coming
Dude!! This is great stuff!! Much appreciate you taking the time to document, and share your knowledge with everyone.
Quick question if you have time. Would it be possible to drive my rc mower manually, from my RC controller, with the cube passive but online, then add my waypoints in the spreadsheet as I progress?
Theoretically, one time cutting manually using the RC transmitter i'd have it mapped, then turn my auto/Man switch to Auto, switching from RC mode to Autopilot, and operate off the GPS.
Oh yea you mentioned distance in the ethernet cable being a consideration...Do you know the max length by chance? Id need to run at least 5' from the rear up to the controller in the front.
Can't thank you enough for your hard work, and allowing everyone to benefit from it! Kudos!
Thanks
Dave
Certainly - you can either download logs from your run and parse through those or (probably easier) just pull up the "data" tab in mission planner while the mower drives around. The current GPS coordinates are at the bottom left side of the map view in that tab, just note them/save them off in a file somewhere and use them to generate your map file.
Have you thought about using two sprockets for each wheel and two bike chains on each side to make tank treads? It may help with off-road traction and any unexpected objects like sticks.
I certainly did consider... I actually wanted to 3d print some treads. I bet bike chains on the ground would pick up sticks etc, but it would be worth a try. If you do try, let me know!
Maybe you mentioned it and i missed it, but does the program have any type of safety stop mechanism, say if the mower has some mechanical failure and wanders off the gps track? Is there a way to program it to stop and/ or shut down? I live near a highway and would want some sort of safety feature that would prevent it from roaming out on the road.
This is a good idea - I talked about it in another comment somewhere in here. Basically I think it should be added - it'd be fairly simple. The cube orange has a arm/disarm switch that plugs into the usb port. That could be wired to "bumpers" on the outside edge of the mower so when they're hit it would disarm and stop moving
Nathan, would there be a benefit in using the Here4 instead of the Here3+
amazing job with this setup! some water proofing could make the electronics last longer, maybe a 3d printed enclosure or anything else which protects from water.
Absolutely. It's on the to do list. I actually had it running last summer when a storm blew in. A drop of water hit the motor control in just the right place and suddenly the right motor was stuck "on". Luckily I was already on the way to turn it off when it started driving off into my field!
@@nathanbuildsdiy stuck on the most stupid thing haha interesting obstacle! hope to see more content related with electronics and hacking
Wow! This is absolutely amazing!!
Glad you think so!
You might throw a solar panel and a charge controller on top to help that battery not drain too low too many times. Or better yet, see if you can add an alternator onto the engine.
A solar panel would charge pretty slowly, but an alternator on the engine is a great idea. adding it to the list of improvements!
Wondering if you could attach an alternator to the gas engine to keep the battery charged while in operation
Yes - a good idea. I'm thinking of that for the next revision.
Does the mower download the RTK updates in real time or are the RTK updates downloaded to your laptop and then transferred to the mower? Great description!
It's near real time, but there is some latency. The laptop does all the communicating with the RTK network because that's done over the internet.
So the drone gets position info, sends that over the 900MHZ radio to the laptop, the laptop talks to the base station over the internet and relays corrections back to the drone over 900MHz. That's all handled in mission planner though, so once you're hooked up to a base station you're good, no need to mess with the details.
fantastic job. I am not at this point yet with my lawnmower- but I hope by fall
Let me know how it goes!
@NathanBuildsDIY I am in canada so will have to check into this. I can see Brandt (deere) has a fixed base nearby.
I'll have to order the controller.
Hi Nathen, Thanks for the awesome project , Right now i am collecting the material to build one. I have a question? Since Motors are 24volt rated why are you using a single battery of 12volt to supply the voltage to dc motors? How is it going to support the 24volt dc motors on 12volt battery?
I originally was going to use 2 12V DC batteries in series for 24V, but when I ran it I found it was plenty fast with 12V and I could charge the second battery while the first one ran, which I liked even more. That's why I ended up using 12V.
Nathan how soon are you coming out with your updated mower. I’m very interested in building one for my grandmother before she passes. She constantly pays people to do her lawn. I’m not sure if you have any contact information for help with my build? I was also wondering do you have to manually start this mower every time it starts? Could you use an electric mower and connect it to the car battery?
Amazing, this is what I want to do. Good job.
Fascinating! Thanks for posting this.
Thanks! I'm working on a new concept, subscribe to catch it when the video comes out.
Your video is AMAZING! Keep producing more videos.
Thanks! I'm working on a good one now but I'm stuck in debug mode trying to get it more reliable before I post.
This is an incredible project. I made one, but mine is remote controlled. Is it possible to use a sabertooth motor controller, or do I need the controller you used? I don't know very much about autonomous. My setup is ready, all I need is the autonomous parts. Thank you for sharing your talent.
I just did a search on sabertooth pinout and it looks like there are 2 main and 2 aux pins on most of them (dual motor controllers anyway). Do you know if the main is a PWM signal (amplitude basically) and the aux is direction? If so it should work with the cube orange/ardupilot just fine.
@@nathanbuildsdiy Thank you for replying. I have to look into the pwm and which output is which. I use the sabertooth with a remote control. However, I want to venture out and learn more by converting it as autonomous. Again, thank you.
Mission Planner and or Cube Orange searches will unveil an amazing amount of info.
Other than CG issues, is there enough traction and torque to climb a 15° hill? Perhaps with more aggressive tires? With your setup, is it effectively 4WD?
Totally 4WD and these motors are very beefy. I had it climb up a tree and flip on its butt before I got the maps correct, so I think if you had enough tread on your tires a 15 degree slope would be fine. Perhaps put weights on the outside hubs (like we do on riding mowers) to prevent rollovers when making turnarounds
Thanks. I am investigating more aggressive wheels for the hills, your amazon link for the wheels go to a 26" wheel. The ones you used look smaller, is that the case? @@nathanbuildsdiy
Great project. Is there any cheaper alternative to orange cube GPS -RTK?
Yes, certainly. Pixhawks (closed source and less quick processors) can be had in kits on amazon. Their GPS's aren't as good as here3 though. I *think* it's possible to get a here3 talking to one of those, but didn't try it myself
Did you make the frame so tall to get the electronics away from the spark plug and coil; or is that an issue? I would like to make a more compact version.
It does isolate the electronics a bit, but I think the electric motors were a bigger issue. I made it that tall because I wasn't sure how the whole thing would turn out when I started :)
I'm betting you could cut down the height and be just fine as long as the gps/compass are far from the electric motors, and as long as you're able to get your mower into the frame (I have to turn mine diagonally to insert it from the back).
One small point. Please sort out the audio stream for future uploads if you are able? At the moment, your voice only plays in my left ear at 26:00
Yeah, I realized that long after posting. It'll be fixed for the future
pretty cool stuff Nathan, keep it up and look forward to future projects. Now a subscriber
Thanks! I'm working on a good one now
Any update on your next generation mower? I am trying to build an autonomous riding lawn tractor which already has alternator but has Ackerman steering instead of skid steering as in your project. Used riding lawn tractors can be bought for a couple hundred dollars. I would love to use your setup with Here3- Ardupilot rather than AGOpenGPS
Hi! Sorry for the delay in response, been busy with family over the holidays. I'm currently working on a version 2 of the weeder and drone sprayers that I built and haven't updated this one yet, but it's my goal to get it this summer. Version 1 is still plugging along, though I had to replace the car batteries powering it
Hit subscribe if you haven't already, then that build will pop up in your feed once I (finally) get to it.
Nice work! I live on 12 acres. This could be really helpful. Have you considered a battery charger (alternator) to keep the batteries full all the time? Thanks for the explanation. Glad I stumbled on this video.
Yes I have, that'll be in the next build (alternator). Thanks!
Pour entraînement pas de problème car en mettant que sur deux roue sa suffit moi c'est le guidage gps qui m'intéresse le reste c'est de la bricole se qui compte c'est le guidaje droit avec changement de voie et arrêt dans la zone déterminer toujour votre vidéo est très instructif
Wonderful stuff, man! Didn't know the open software was so good now. Even training the AI for your freznel weed killer seems to be approaching dummyproof. The only upgrades I would make are:
Weld on a sorta cattle-guard to push away the big rocks and branches ..
Larger tires just for traction, because i'm building a chariot behind mine to pull me around the property like a retired drunken Spartacus.
Give your map to a drone to do a quick scouting ahead..
Honestly forget even the mower.. if one could just come up with a pathfinding diy platform rugged enough to carry a couple of chainsaws and a gas can/generator water/lunch/guns .. There's nothing worse than losing an hour+ of daylight just because you forgot/need a pair of needle nose.
It's gonna break, of course, like everything on farm will.. I'd much rather have to deal with a broken or jammed up chain than dealing with anything like tank treads unless the property absolutely requires it.
I love the project! A suggestion would be that during the voiceover portions (like at 4:33), your voice only comes out of my left headphone and it's a little bit annoying
You're absolutely right. I didn't realize it was mono audio on the voiceover until after I had published. TH-cam won't let me upload a new video without deleting this one. All future vids will be stereo audio.
I'm doing my build now and looked at the video and schematic again closely. For Aux out 5 & 6, what did you configure in Ardupilot for this setup (wheel direction) ? I'm using the bigger Cytron MDDS30 and it is slightly different. Also what did you set for MOT_PWM_TYPE? I'm using wheel chair motors for my setup.
Did you figure it out? I'm using the same setup as him but I can't get reverse working.
Please let me know were and how to download the software for gps controller. It is Wonderful Assembling and you are genius. Please let me know if the gps section we can do it cheaper from china cart gps. Thanks, I love it.
Thanks for the comments!
Here's where to get mission planner, the software for the base station (it's installed on your laptop):
ardupilot.org/planner/docs/mission-planner-installation.html
Here's where to get ardurover firmware - it's installed on the cube orange (or pixhawk) autopilot:
ardupilot.org/rover/docs/common-downloads_firmware.html
There's good documentation about getting these set up here:
ardupilot.org/rover/docs/apmrover-setup.html
This is so sweet. Well presented.
Thanks! I'm working on a new concept, hit subscribe to catch it when the video comes out
my left ear really enjoyed this video
LOL, well played. I didn't realize the audio was mono when I uploaded and I can't upload the stereo version without deleting this video! I'll fix for future videos
a 50w solar panel with a solar charge controller, could be used to protect the electronics from the sun and recharge the battery.
Exactly - next revision I think will have that.
question, you show 24v ebike motors but they are hooked up to a 12v car battery. how does that work?
They run at a lower speed basically. I tried running at 24V initially but the whole thing moved too fast for the mower to keep up. When I tried at 12v it was happy.
@@nathanbuildsdiy were you not able to control the speed on 24v within mission planner? appreciate the info, ive ordered the parts to build my own.
That's great! Let me know how the build goes, if you've got any questions post a new comment (I always see the new comments, sometimes miss replies). We can connect outside TH-cam for any help/questions.
As for 24V - yes, you totally can set the max speed in mission planner (it's on the basic tuning page to the right of the P I D settings I think). When I started 24V testing I didn't realize that was there. I tested 12V and it was a speed I liked so I kept it and got into the habit of charging 1 battery while the other operates, so I've always got one ready to roll. Much later I realized I could have stuck with 24V and set a max speed but I had already developed a routine.
If you're going to build here's my list of ideas for the next version, pulled from the comments here, if you want to try any of them (no need, but I'm thinking I'll add next):
- add a front bumper to stop if we hit something (connect to the arm/disarm button that goes to the USB on cube orange)
- Add a small alternator on mower to rob a bit of power and charge up battery
- Add a chain tensioner.
- Put a pin in the axle so it can be removed rather than welding that inside washer to the axle
- 2 gps for yaw/no need for compass - But man... this is more $ and I just can't pull the trigger
I built a robot similar to this a while ago and found that the GPS resolution was not fine enough. Non military GPS can have a variance of up to 3 feet. How did you prevent the mower from running over Grandmas flower bed?
GPS RTK - it's base station corrections which get you to sub centimeter accuracy. I discuss how to set that up in misison planner with the here3+ GPS in the middle of the video. That's also the reason for the more expensive GPS unit, it can do RTK.
Is there a reason you used the 24v motors? Do they still work reliably on the 12v battery?
I planned to use 24v (2 12 in series) but realized it was a good speed with 12V and I could charge the other battery while one was running. Now I'm working on an alternator/regulator to recharge the battery while it runs and that's a 12V system so I'll stick with that.
U can use roveboard engine.
Yeah, I thought about doing that and couldn't find any used ones for cheap. You want to have rpm's fairly low - these motors do the gearing down inside the enging housing - something like 100-200RPM depending on the size of tires you have, and plenty of torque at the low speeds.
Nice.... But is it cheaper than buying one off the shelf?
Great video, thanks for sharing
Can you build a RC FPV tank using these same components?
So you run the 24 volt motors on 12 volts ?
I am a new follower of your channel
Greetings from Sweden
Welcome! I just put out a video on a robotic weeder, check it out.
To get Aux 5 & 6 to work for directional control, did you have to set any parameters? I have the same setup as you. Pin outs on the controller and cube have been double checked and the cabling is good. However I can not get the motors to change direction. Any suggestions?
Sorry for the delay! Been busy with kids starting school.
Lets check the settings -
Initial setup tab -> Mandatory hardware -> servo output -> throttle left/right are channels 1/3, correct?
Config tab -> basic -> Brushed with Relay
You're connecting to the AUX 5/6, not the main out 5/6, correct?
Next go to the Config tab, under advanced (I think?) and search on the word "aux", it should pop up what those are set to. Next up... I suppose check with a multimeter connected to the output pins to see if aux 5/6 are actually changing voltage.
@@nathanbuildsdiy thanks for the reply. I will run through your suggestions to see what happens. If I select ‘brushed BiPolar’ then I can get the mower to turn. If I select brushed with relay then when I command a turn both wheels still go forward. I do have it set up for rc control also, that might be a factor also. Anyway, I was just looking for a quick fix, not product support. I have the winter to get the bugs out. If you would like here is a link to my video of the mower, I referenced your channel in the video and included a link in the description. th-cam.com/video/mMNRO0pRbzY/w-d-xo.htmlsi=ES4XU3riRxZxvhAf
@timrit3817 did you figure it out?
Please upgrade it with junction boxes for electronics. THX!
Kudos for this build. I'm impressed how this works. But I would never run a live mower withoud additional sensors, that avoid objects PLUS constant observation. So I wouldn't benefit from this build 😞
Totally understand. I'm fortunate that the farm is quite isolated and surrounded by trees and fence. I'm the only person or pet around when its running and I work nearby so i can keep an eye on it.
Would love to see you put a solar panel on that. Electric start mower. Battery dies it cuts mower charges to a specified point cranks up the mower and keeps going. Would be the next step to a totally solar autonomous mower just need to go lithium and electric mower motor. It would be a mechanical goat.
I like the idea, but I bet the size of the solar panel needed to meet the power demands would be pretty large. I think next build I'm going to add a small alternator to the mower to power the battery
Amazing channel.
Amazing content.
Please keep going!
Thank you! Will do! I've got some good projects in mind for this summer, check back!
How are 180 degree turns handled? Is that an algorithm in the software or done with waypoints?
Yeah, all that is handled in ardupilot. I set waypoint "size" as small as possible, so it doesn't try to make a big U turn but instead pivots in place.
Amazing work, thanks for sharing!
Thanks for watching!
2 kewl. I have built a few RC mowers and would like one to be autonomous. Does this require any kind of ongoing subscription fee or is it a one time cost?
No subscription cost to connect to the base stations here in Iowa. Farmers use them all the time for their combines/tractors so it may be different in your area, but all systems I've seen in surrounding states are also free
just a question would larger 26" wheels work as well
Yes, but it'd put the whole things up a lot higher off the ground so you'd need to hang the mower down from the frame to keep it near the grass.
Good work. Lots of useful information👍
Glad it was helpful!
This project is awesome
pretty big motors for moving that, would you say they're overkill? I see other builds using 12v 60watt powers, massively less
Yeah, they are large, but were a good price on Amazon so I went with them, plus the frame and car battery are pretty heavy, so the extra power isn't for naught
Awesome video mate love it :)
I'm curious to know more about your shed structure!
It's a barn I built myself. The posts and rafters were all cut from trees that I removed from the surrounding fields (they're a type of weed tree around here, always going where they shouldn't). The barn boards were salvaged from my grandma's barn that fell in a big storm about 15 years ago. They're about 140 years old!
I was actually thinking of a video on how to build that, I'll make it now!
@@nathanbuildsdiy That would be great I'll watch it !
solar panel on top to recharge the bike motors and you'd be golden!
I like the idea, but I bet the size of the solar panel needed to meet the power demands would be pretty large. I think next build I'm going to add a small alternator to the mower to power the battery
HA how nice!
How easy can it be.
Nathan was the name? well done man!
healthy and kind regards from the Netherlands!
Rob
Thanks very much! I really enjoy tinkering to see what I can accomplish and this one does a pretty good job.
thanks for this very well explained