This is my employer's bread and butter for many decades: https://www.ravenind.com/products/applications-booms
And yes, it's a lot of emphasis on autonomy now, but you'd be amazed at what just having better spray nozzles can do?
Cash crop farming is an incredibly capital intensive operation. Robots don't fix that.
In my hydroponics system, spray nozzles are one of the more expensive (for the size and purpose) components and they fail the most often. You’d think a brass nozzle would last years, but it’s going to need to be disassembled and cleaned before a year is up. Now imagine a huge tractor with high pressure bars that are lined with dozens of nozzles operating in a dirty environment and moving huge volumes of fluid. You’re going to get clogs, you’re going to have to clean them, and the application will be non-ideal quite often if you don’t. And that’s just the nozzles.
This stuff is so much harder and more complicated than it seems on the surface. Adding automation almost makes it harder because you really do want a human observing the system and ensuring it’s actually working.
And it's not just fertilizer, but other sprays, too.
Here, this looks thoroughly boring, but this guy is going over configuring one of our systems for fertilizer application... You can get an idea of some of the variables involved, if you can make it through :-)
But I think we're kind of in a hiring freeze. I think I just barely squeaked in before that.
The above is not a complete list of all the nozzle considerations I know of (I work for John Deere, but I don't speak for them), it is what I can remember in a couple minutes.
Mind telling your bosses to fuck off and let farmers repair their own equipment, please?
Or does that make absolutely no sense? I’m not sure how you can reliably measure fluid properties in order to get consistent spray patterns. It seems like it would be useful here, though.
You can tell I’m not an engineer :)
Usually when someone says "AI will save the world because" it is followed by a bunch of naive hyperbole.
However, you are 100% right. A few robust robots with an AI to spot pests and poor nutrition could decimate input costs for a farm and solve several environmental issues in one go.
It's the simple reason why a well-tended garden produces better produce than a corporate farm. There is simply more effort per plant in a garden. Tractors are currently too dumb to treat plants individually. They require plants to emerge, grow, harvest at the same time, and in the same way. Current modalities for pests/nutrition treat plants at the field level, not at the individual level. This creates lots of waste and externalities.
If there were a robustly built robot smart enough to deal with plant issues in and individual way then there is a huge slice of efficiency that could be gained. A field with the productivity of a garden. People have been trying to create this for a decade now and the limiting factor is the software. Maybe with the latest round of AI advances that will change.
Note however that modern industrial agriculture uses fertilizer fairly inefficiently, and a large fraction just runs off the fields into waterways. Here is where the AI robots could make a big difference, by crawling up and down farm fields and selectively applying fertilizer to individual plants as needed (while also removing weeds and treating pest infestations). Of course, this doesn't make agrichemical producers all that happy as it could greatly reduce demand.