Autonomous Gardening

>>1212789
see
>>>/out/1062604

File: pivotcenter.jpg (26KB, 400x267px) Image search: [Google]

26KB, 400x267px

>>1212789
I've actually thought about his quite a bit. You know those center pivot irrigators, make one like a gantry crain with a robotic arm that travels up and down. Could even put a plow attachment and whatnot on it. Would need to be beefier than a regular aluminum center pivot one though. That was my idea anyhooz. Prolly a pipe dream.

>>1212789
Weeding is unnecessary in a hydroponic setup and you'll have more fun automating that than a normal garden.

Czech it
https://www.theguardian.com/environment/2016/feb/01/japanese-firm-to-open-worlds-first-robot-run-farm

Everything I've seen on the /diy/ scale was a design much like what you posted, which is the worst one.
>big, expensive gantry moving over straight rails
>rails are usually short
>each planter needs its own gantry, tools, et cetera
You need either a system that uses a single rail with an arm that reaches over the planter, or an unbound wheeled vehicle. The former is more for hydroponics/aquaponics, the latter for regular outdoor farming. The latter is also being developed and applied already, so if you want something new, you'd want to come up with either a smaller wheeled machine that can navigate personal gardens and care for the crops - will probably need to develop a local positioning system for it - or a rail hydroponics system that does't suck.

>I have no clear understanding of farming and what tasks would better be suited for automation
Weeding is not. Harvesting requires the machine to be strong, depending on the crop you need to dig the ground or cut the plant. Planting, fertilization, watering, soil monitoring, packaging and maybe heat protection can be automated easily enough.

>>1212789
overcomplication of stuff that is suposed to be simple. Why would you need an robotic arm in a planter-sized field? The only purpose I can see is to harvest squishy things like fruits. But that would also require a system to detect said fruits.

>>1213168
Also, fertilization is done only a couple times during most plants lifecycle, you could do it manually. The things that are very good to be controlled by machines are
>temperature
>air humidity
>water
>sun exposure
>lights

>>1213168

The goal is to build something that will operate autonomously and make something that today is very labor intensive, require no man power.


Also making it autonomous, improves efficiency and care of the farm, requires less use of water, and no use of pesticides.

This is the future of farming. Of course I want to build something.

>>1213214
Yes but for small scale things it isn't labor intensive. What you want is a farm with farm equipment. Most expensive combine harvesters and some tractors drive themselves.
There is no automatic system that takes care of the entire cycle. Making a small scale one would be pointless because small scale farming requires not so much effort. And making a large scale one so far hasn't been done. Why don't you do a automatic greenhouse or a aqua/hydroponic system?

>>1213220
You could do a system that takes care of water, sunlight and temperature for example. I can already imagine a couple things you could do with PICs or MSP430 to control dripping rates, shades, fans etc

>>1213221
the good thing about using embbed systems is that you could code them and that code would be usable to any instalation. You will only need to buy shit then. And you could also hook up some transducers and record information about your plantation.

>>1213220

There is a growing market for small scale "back-yard" farmers and people making an income by selling at farmer's markets and local restaurants.

Imagine utilizing every suburban space possible.

>>1213227
Yeah I know but that kind of thing require almost 0 labor. I'm one myself (or was as of 4 months ago). You can automate ligthining, temp, humidity, water and nutrients. But packaging, harvest etc? Be my guest

>>1213232
https://www.youtube.com/watch?v=UlaNDm88yZo

If you can do this, ok then go ahead

>>1213092

You can get gps modules that are accurate to about 0.5cm when within 500m of a base station for ~$400 (includes base station).

>>1213238
You can, but you can also usually go and fuck yourself with them, which is what I did. RTK needs signal lock from over a dozen satellites for that precision, anything and everything breaks the lock too. Though I got the standard GPS variant for $50 each, there are now some GPS-GLONASS models for more shekels, but I'm not too willing to test again.
Instead of that I've been trying ultrasound for local positioning using cheap low-power satellites, which is quite promising, but I haven't had time lately to work on it. This was for an autonomous small mower, but in principle a watering-fertilizing robot could use the same technology.

I think the best of the best would be optical positioning, using landmarks/markers, but that is a bit too complicated a project to take on alone.

File: IMG_20170721_145747_HDR.jpg (3MB, 4608x3456px) Image search: [Google]

3MB, 4608x3456px

>>1213092
Not OP but this thread perked my interest. I've been fiddling with the autonomous farm bot concept for a while. Here are some sketches with bad handwriting. They are kinda like electric wheel chairs with forklifts.

File: IMG_20170721_145754_HDR.jpg (3MB, 4608x3456px) Image search: [Google]

3MB, 4608x3456px

>>1213314
I also really like that circular grow area with a rotating arm above it, they had something like that in Tropico 4. It would plow, seed and water fields. Was a neat idea.

>>1213318
>>1213314
now code them

File: garden_robot.jpg (337KB, 1024x961px) Image search: [Google]

337KB, 1024x961px

>>1213353

>>1213035
You'd have to make sure the hinge is stronger than the core of a neutron star, and you'd have to have a thick steel pole and a tractor or something to provide the most torque.

Seems pretty unnecessary desu

If you make a greenhouse, weeds and shit wouldn't be a problem. You said you had resources for seed planting, another suggestion I had was to make an overhead watering system that releases based on a timer. I mean fully automating it would be a pain in the ass, but planting and watering being covered would take out most of the work just leaving harvest. Maybe put the plants on a conveyer belt in the greenhouse to even automate moving shit into it.

My set up would be outside the greenhouse where starting materials are, a loading area for conveyer belt, load pots with soil in it and have the seed rig put seeds in while you are loading another.

>>1213365
gosh i wish i had one atom of artistic ability in my entire body.

>>1213232

OP Here

yes I know.. packaging and harvest are probably the hardest. I am not saying it will be 100% autonomous but using computer vision (Something like OpenCV) + AI (Run a dataset of whatever I want identified from photos from google but ones I take as well), or other sensors that could detect ripness and the vegetables more efficiently maybe I can make something efficient for harvesting.

I am thinking of a small robotic arm with an adaptive gripper. Maybe even increase the surface/width of it to improve grip.

https://www.youtube.com/watch?v=90cXfaFM4O8
About packaging, After harvesting it would be the easiest part of the project to place the produce in a fridge into different crates.

>>1213314
Deere used to host collegiate autonomous lawn mowing competitions.
This is pretty much the same thing.