I am excited about your project! Small suggestion for a video editing tweak - your intro music is much much louder than the voice that comes afterwards. Thank you for all the work you've put into this.
That's crazy, you make an outline and any areas not to go and it pretty much makes G-code for the mowers path. This seems very parallel to a computer controlled plotter, laser, or even a 3D printer nozzle moving on a first layer of a print. Really cool idea!
A colleague and me talked about doing something like this for years but never had the time... We were into GPS, camera etc... But having it learn by using a controller is fast and smart. Hope this pans out!
@@pietheijn-vo1gt Indeed... I'm glad some people have/take the time and make something like this happen. It feels like camera and image recognition to avoid kids toys in the yard etc could be a cool future iteration.
@@DanielLiljeberg we're in the discord talking about this constantly. We are looking at mmwave radar, ToF depth sensors and camera's for object recognition. There are many paths to take here
@@pietheijn-vo1gt Sounds awesome... I see the solution is based on a custom board that is not available for purchase yet. When it is I'll be sure to buy one. My wife has been asking me to get a robotic lawnmower for years (I guess her other robotic one... me... isn't good enough)... But I think they all suck. Burry some cable just to enable them to run around like a drunk person... Xiaomi showed of one that was base don lidarr and cameras I think that sounded like it could be cool. But I haven't seen it actually available to the public yet.
@@pietheijn-vo1gt Also, not sure how accurate the GPS is, but given that the mower i trained using the controller it might not be an issue. I remember that we were thinking of using something like low power bluetooth or something like that from a few placements around the yard to get very accurate positioning.
Great video and nice work! It would be great if you did a diy build and showed us the full process of building one. Will be following your channel for updates!
This is a great project, you'd think by now there'd be a commercial automated lawn mower that works great. Maybe you could even sell this as a kit! Or kickstart an improved version with a single main board.
@@RizZRizZ- the complexity to make a controlled path is way greater than letting it bounce around randomly and there will not be any noticable difference in the end except for the higher price. running the same path over and over will also be very visible in the grass and it will be ridiculous to create setup with a controlled random path rather than just letting it bounce around as it already does
Yes! I have been thinking of doing this exact same project for *years*, but have never gotten around to it. One use case for "structured mowing" is challenging hills. Let the mower climb on lesser incline and let it descend down a steeper hill. That way it won't get stuck and dig a hole with the wheels. You could draw routes and perimeters on a map in your browser. Bring it with you on your vacation without having to install stuff (except a RTK base station). Move it it from place to place and have perfect fixed object avoidance. Now I just need to find the time to follow your plans which is a much lesser problem than starting from scratch :) I have played around with that same Ardusimple GPS before. Stupid accurate! And in my setup I had a dual gps rover setup that gives perfect heading even at stand still.
Really great work! I love how you can tell how much attention to detail and effort went into the project. Looking forward to seeing more content from you!
This is my dream! I also want to create such robot by myself because I kinda tired to mow the grass. But also for me one of the important feature would be in such robot to have storage for grass. Sometimes you use it sometimes not
Awesome work, and brilliant idea. I had the same thought when I started to look for a lawn-mower robot: they're basically dumb! So I'm not buying one until they either get smarter or drastically cheaper. They basically all require some wiring all around the area and can't really jump by themselves from one area to another (I, for example, have two areas to mow in my garden that are separated by a paved area). I always wondered why lawn-mower robot manufacturers struggle that much in making intelligent devices whereas vacuum robots like the one I own for already 6 years (Xiaomi Roborock S5) is literally painting a map of my house and discovering it by itself. Anyway, I'm really glad that some smart project like yours starts moving things forward!
Yes, that was my point. It's either dumb or too expansive, so not really interesting right now. With this project, the poster's intention is to make it both smart and cheap, which is definitely the way to go.
Looks like loads of companies have taken your idea and an hour flogging it around the world. I’ve had a robotic lawnmower for nearly 9 years I’ve been waiting for GPS. I’ve just come across your RU-vid channel at the same time and seeing lots of companies launch their GPS-based robotic lawnmowers there obviously using your technology. Well done.
Can I make a suggestion/request? Can you have different mowing "patterns"? Something where after it goes back and forth, maybe it could do diagonal lines to ensure full coverage.
Your mowing simulator looks really realistic. Really cool how you program the thing. Not having to dig a bunch of trenches I nthe ground like the old style that followed a line under the ground kind of an electric fence. Hope this picks up and people can help you improve it since it's n open project.
Amazing stuff. I should have seen it coming, open-source is definitely the fastest way to get these things working without boundary/guide wires. The big vendors will not have the ability to put in the effort like an open-source community will. But they could and should join forces with the community and implement this solution into their product lines, rather than waste their time trying to develop their own from scratch. The idea of automower's coming with onboard API's to tie into with the open-source control software is kinda cool.
Totally awesome. I look forward to your updates! I can’t spend $900ish for the shell and motors, so I’ll be looking around for a 3d printable version of a suitable shell. Again, great job!
I had i project for my engineering program in high school and I literally drafted the exact thing but didn’t pursue it because the rest of my team thought it wouldn’t be a good project
Yes please. This is great. I liked the idea of automated mowing but I did not like the fact that you had to leave it out all week while it randomly mowed. This is perfect, I have have it mow for a hour or two while the kids and dog are inside. RTK GPS is amazing!
Awesome! Next step: smart vaccum robots! :D Thank you for making this open source. Not only does it seem much more efficient and logical it also seems a lot safer in terms of data protection, when you don't have to use or upload any layouts to a shady "smart" app.
there are already vacuum robots that are not randomly working and the entry price is also reasonable. Look on the xiaomi roborock s1 for example. The problem with mowers is they are all just random bumping robots. It would be interesting to gut a xiaomi vacuum as it has lidar and put it in a robot mower shell
A few cars use lidar already - and not having any problems. Of cause they use way more complex lidar sensors which cost >10000€... The vac robots just use a slow spinning single laser in one horizontal plane. Way easier to do - and cheaper. I guess this kind of sensor you could buy for a few bucks. But then you need a mapping solution which is not easy to do - but I am sure openly available (saw a yt video where the guy build a walking robot with this kind of automapping solution). The differential gps is a nice idea and probably the easiest to implement. This will not work inside the house though (GPS signal is blocked inside a house) - that is the reason why they went with the lidar approach...
@@DanielBulyovcsity Yes, I wrote that these systems you could just buy. Open source software available. It is not much more than a spinning mirror with a laser and a diode... But in bright sunlight probably not working well. And it is only giving you a single plane of the boundaries. Of you have a hugh yard with no fence around - the laser will not see anything. The mapper algorithm needs obstacles to "attach" to and create a map. My ecovacs sometimes looses this even in a crowded space inside the house...
Nice Video of a very interesting project. But to give you a hint about improving your video quality: The Volume levels shifting very drastically. 1. The music at the start is way to loud 2. The first explanation is a good amount of volume 3. The next explanation phase is too low So watchers are need to adjust the volume back and forth, that takes attention away from your content ❤
Awesome project! You should do more videos on how some of the modules were built (the mechanics / electronics / programming) in some kind of tutorial fashion to get more exposure for the product
I need to build one of these for my side yard. I'm tired of the neighbors bothering me when I just want to listen to podcasts and mow as fast as possible.
Wow, a few months ago I was thinking about how terrible robotic mowers are right now. Closed-source hardware and software are far slower than open source for niche electronics like this, and you have ZERO control as the owner. The only innovations they'll create are the ones they can charge you for, and why update last year's firmware when you can release a new model for more money? You're doing the Lord's work here!
Amazing work! Love the attention to detail and the whole open-source approach. Probably will build one this summer if I manage to save up the resources to buy everything needed :P My dad (me as well, probably) would freak out if he saw this thing in real life :D
I'd like one that has 'music mode', which enables it to follow me around the garden playing tunes from an integrated speaker. A cooler box trailer might be useful too.
I pay a person $35 each vist to mow my lawn because I am old and can't do it myself. But I am a maker. During the winter, I can save that $70-$105 dollars per month to buy the parts needed to build the RTK base station then make my own robot mower. See Andreas Haeuser's RU-vid video to make the mower, just forget about the radio part that detects the "invisible fence" wire around the perimeter. Then, in the spring, I can cancel my lawn service and let this mower pay for itself WITH CONSTANT SUPERVISION so the kids in the neighborhood don't touch the cute robot mower thingy!
Today I learned about GPS RTK, which apparently gives you one centimeter accuracy once a second for about $250, I had no idea the system existed, or was this affordable for its accuracy
@@ClemensElflein need for a base station and still depend on a satellite network (Kessler syndrome might kick in sooner than we expect) ... would a set of BLE beacons provide the same accuracy without external dependencies and even cheaper?
2 года назад
@@ClemensElflein Is this base shareable over the internet? They should be valid for a large:ish distance, right?
@ I just found this video but I have been experimenting with RTK for a few years. The department of transportation for my state (Iowa) maintains a base station network that covers the entire state intended for farmers to use. I was able to get access just for asking...no charge. You might want to check your state website.
2 года назад
@@kbouwman64 I checked this up. I don't live in the US and sadly here in Sweden we do have a network but it costs a lot of money, well honestly not a "lot" lot, but just a bit too much for a mower maybe!
Hi Clemens Ein Open mover brauchen alle die einen Garten haben. Ich würde mich trauen es anzugreifen, habe aber keine Programmierungs Erfahrung. Dennoch werde ich mich über die kommende Updates freuen, und Mal schauen, ob ich die Teile kaufen kann, wenn deine Videos dar sind. Super Projekt!
Nice project. I know a lot of of people also try to build robotic mowers with hoverboard parts. You could also drive a hoverboard mainboard from the raspberry.
Great job👍 Hope you stay tuned and let us continue to participate in your project. I would definitely look forward to a video series about it. Thanks in advance.
Holy S, this is great! Even though I don't even have a garden :D Maybe you could get in touch with some front end dev to make an app for that, that would make it easier to monitor and setup. Let the Open Source Fun begin!
The latest-long signal using Single GPS on mower will shift a lot depending on the whether. I’d use multiple gps receiver to measure the relative location to a fixed gps.
have been looking into robot mowers the last weak and felt the need to just diy one myself, now I see this video and yea... I guess that is what I was searching for ;) Awesome project, will look into it and if I build one start contributing :D Thank you for making this video 👍
I have never understood why the mowers of big companies don't work exactly like this. When I first heard of these robotic mowers, I assumed they would be working exactly like this and only later found out that basically a fourth grader programmed them. Absolutely laughable. Absolutely incredible that it took like 10 years and an open source project to make robotic mowers make any kind of sense.
Hi there, I sell Husqvarna Automower, there is a big reason why we dont use GPS: The moment it fails for seconds or 15 cm. It can go directly into the pool or the street directly to traffic. Also some gardens, most of them to be honest, are highly complex, where GPS can and will miss into the bed of flowers or so. I think our system of random movement and perimeter works very well.
@@lucio9601 Well, GPS can be combined with accelerometers and gyroscopes to make them much more reliable. That would bridge any momentary loss of GPS signal. It should be no problem for a company like Husqvarna to develop such a system. The argument about complexity is definitely a valid one, but also means, that through random movement the mower might have "blind spots" where it never mowes. I think it should be possible to make a system that can mow a complex garden with an efficient pattern. 3D printers basically solve the same problem.
@@Blockbuster2033 the thing is, most people dont want patterns in their lawns. The random system in Husqvarna gives a full cut in a delimited area between 2 a 3 days of work. (Be it a garden with a pre cut). People just want to set the robot to work and see results in a good time, nothing more. You can make complex models with gps help and area control like Husqvarna 450X, and that's it, they are "simple" devices.
@@lucio9601 it would be trivial to design the system to automatically stop moving if it cannot receive location data -- your argument sounds like fear mongering from a salesman :)
@@PabloEdvardo I think there were auto mowers that worked with gps before.. but they all fail to meet the clients desires. As I said before, a bad mistake and the whole thinks falls down into the pool
This is brilliant! Current mowers are stupid. Please offer a kit, make an app, etc. I’d happily pay 1500 EUR for a good, serviceable (and extendable?) mower with great software.
That is a nice project, the robot learns from example I have some projects that do the same where I had an option for motion-by-queue actions and all AI learning but seeing it running in the backyard really creates a value project.
Great work! I think it should integrate the cable sensing too. For example: I have a automower 310, I have the cable already in, like the concept, but would love optimized mowing in our complicated garden. Honestly dont want to trust ONLY in gps if the cable is already there anyways. There are so many automowers outthere, self repaired, little upgraded ones too, there is a big market for you outthere with the huqvarna automowers. I'd love an upgrade kit ;) Would pay I guess 300-500 for it.
Very cool! I worked on building my own custom autonomous lawn mower 5 years ago but eventually lost my ambition. I was hoping to base it on an Intel Realsense 3D camera sensor. Very cool that CM level accuracy GPS is cheap enough these days! I will probably give this a try, although my yard is much bigger, so I will want to explore whether there are bigger off the shelf platforms.
@@janramonmartin I've come to the same conclusion after doing my own research for a few days. There are two ways to make RTK work: 1) Give the robot internet access so it can stream RTCM correctionsl from a base station that is hopefully in your city. This will still cost ~$500, not including the cost of the Wifi network. I'm not a fan of this solution because it depends on having your yard blanketed with reliable Wifi. I want my robot to work offline. Option (2) is to build your own base station and somehow wirelessly communicate the correction data. This will cost 2x$500 + your wireless communication solution.
@@chrisraper6181 im looking further into lidar and vslam, when you look at these applications it is largely irrelevant where on earth the robot is, so it makes sense to look at specialized localization tech (i hope..)
@@janramonmartin I've decided to give a depth camera another try using the RealSense D455. It is sort of like short range lidar (10m max) plus vision. Not sure what SLAM algorithm I'll try yet....
Wow. I just found this. THANK YOU. I also just got an older robotic mower that runs great. Has 3 blades and powered wheels and mows just fine on random pattern and wire perimeter. This one is so old, it doesn't have a base station and automatic charging. Would love to improve it to the latest GPS RTK application. I'll check out your links. But I would very much like it, if you would do a complete video on your build. Starting with the mower and how you have installed everything. Especially the programming.
Amazing!! I've been wanting to build a robot mower for awhile now and planned to use GPS like you did. Thank you for the work you put in to this project and making it open source. I will definitely check it out!! My other complaint is that commercially available mowers are built to cover city sized lawns, leaving the market for acreages untapped. I own 2.25 acres. There's no way any current robot mower is going to cover that in a reasonable amount of time. Lastly in addition to a robot mower, I could see this used in snow blower applications as well, maybe even make it modular with different attachments so you don't have to build a 2nd bot.
I think an improvement would be to have the zigzag pattern angle change randomly between sessions. otherwise the lawn will have zones that are more worn out by the robot wheels.
I've been mowing my lawn for almost 50 years with push mowers that are much heavier than this one. Not a single wheel path, or even an impression is to be found.
@@snakedoktor6020 You also aren't a robot that generates the exact same path plan every time you mow, and follows it with a GPS module to ensure consistency.
@@Phoen1x883 You are correct. I am not a robot. But, I cut the grass in the same pattern, from the same starting point, every time. There is just not enough weight, or enough friction generated, for this to be a concern. And if it were, just let the mower program a new route each summer.
Wow I love this. I have the exact same yardforce mower. I'd love to apply your brilliant mods. I don't suppose you are willing to put together an "upgrade" kit to sell to people like me 🤔
Thank you for your comment! I'm glad that you like my project. At the moment I'm not able to sell anything, since this is "just" a prototype of mine. I basically wanted to share the project and didn't anticipate the large response. Starting a hardware company will be a lot of work and I probably won't be able to do it in a short amount of time. You can however follow me here and / or on Discord for the latest news 🙂
I would buy one. Now if it can have a roller behind it to help with stripes then take my money
2 года назад
I do like the teach-in function and the RTK GPS functionality, but I do want you to reconsider the random pattern, they are using it for a reason, if you read patent papers from Husqvarna about the issue of tracks in the lawn you will get why they using random patterns - or you will understand it when the season is over! Never the less I really liked your project! Keep it up!
Excellent project, great design, and very practical. I absolutely love it, and I'm following to keep up with any updates and additional info. I really would like to build this. Couple thoughts/questions: 1. Have you thought about making the added circuit boards as a compact add-on component that would simply mount to the top and have an external wire harness that would plug into the mower? This would make it more compatible with other brands of mowers, or even other types of vehicles (snow plow RC car?). This would also make it much simpler to package these components as a single unit for sale, and I'd be happy to help you out with that venture. 2. Have you looked into incorporating LiDAR for object detection? Or perhaps using an existing product that uses LiDAR (such as a Wyze robot vacuum) and incorporating other functionality into that? 3. What does the UI look like for mapping? Could you program no-go zones in the mapping? Any thoughts on an android app UI so it could be controlled via wifi from anywhere?
None of these are important since he uses a 500€+ robot mower that already allows programming. I don't see the point of this project other than "I learned to control an already drone mower with an arduino and a RPi".
@@TrasherBiner Weeeeell, not quite. He's given a low tier, off the shelf lawn mower drone pathfinding and mapping capabilities superior to most top models (which in turn cost more than 4x the price of this base model). Even considering the cost of the additional hardware, that is quite a feat! Plus, the (claimed) capability of mapping multiple, seperate (!) areas makes the setup quite unique...
I love people that ask questions when they haven’t watched the video. #1 was partly answered in the video and #3 was shown explicitly in the video. Adding LIDAR for object detection would only be necessary for objects that weren’t mapped, almost always a moveable object unless the city installs a sign in your front yard or you plant a tree. Those things won’t happen very often and you would remap the area most likely. Or just let the mower bump into the new thing, stop, and go around it every time it mows the lawn. Not a big deal.
In the 1950s through the 1970s they had a mechanical robot mower. It would mow in a square pattern starting at the center and working it’s way out word with mechanical clicks. It was way better than the robot mowers from the recent past. Find me a robotic Mower that’s better than a 1950s robotic mower!!
I love this project, finally a proper controlled mower. RTK GPS modules are pretty pricey. Almost as much as the Yardforce mower in the first place, but can't really complain for the amount of accuracy you get. I'm really surprised that mower has sensored brushless motors for the wheels and blade at that price, it's very interesting project.
Great work! I bought a Bosch mower last year and it SUCKED! Took me 2 days to lay the wire and it had to mow the lawn 10 times! Until it finally „understood“ the lawn. Please license the software to Bosch so they have it in their next updates! :)
At least husqvarna has proper algorithims/GPS perimiter detecton (their higher end mowers)/great wire detection. My 450x even knows when it didn't mow a part of my jard for a while and it just focuses on that part that session.
