Experiences with FarmBot or similar gardening robots?
This is just a random thought I had. I don't do gardening currently and not looking for advice per se. Just thinking about how the physical world feels far behind in terms of automation compared to the digital world, and wondering what kind of possibilities are out there. I was wondering how close we are to having consumer-form-factor robots to help with various things, and growing food is a natural starting place.
I was imagining what kind of robots are needed to deal with a garden—assuming a house with a plot of land suitable for a large garden—with tasks like:
- Fetching water, either from plumbed water or a natural water source
- Getting seeds from somewhere. Maybe online shopping and then the robot knowing how to open the box. (Probably not by identifying existing plants and picking/stealing them.)
- Planting the seeds in the right place
- Watering the plants regularly
- Maintaining temperature and sun exposure
- Digging up the plant and bringing it indoors so I can inspect or smell it without having to go outside. Then replanting it safely.
- Determining when food is ripe, picking it, reusing the seeds
- Washing and cooking it
It feels like a lot of these are already available off-the-shelf today. I searched and there is a project which I hadn't heard of before called FarmBot which seems neat and geared toward enthusiasts ("prosumers") and education, and includes open source hardware and software. To be clear I'm not affiliated with them in any way.
FarmBot probably handles a lot of the important parts of gardening, but I'm sure it doesn't handle everything on my list. How far are we from a 100% automated experience?
Other than that there was some recent marketing around cheap robots like LeRobot by HuggingFace (the company where basically all the open-weight AI models are hosted). It has nothing to do with farming except that they have one shaped like a hand, so it could probably be programmed to grasp and move things around.
Sorry for the rambling post. Really curious to hear if anyone else has gone into robotics and interested in hearing your experiences and also other resources on what state-of-the-art looks like. Also I bet a lot of this is solved in proprietary solutions and by Big Agriculture, but right now I'm more curious on the consumer-grade level.
I worked on a commercial vegetable-breeding farm that did some (very, very expensive) experimentation with robotics, primarily for weeding. It did not go well.
The weeding robot was highly ineffective because it could not actually differentiate weed seedlings from crop seedlings, it could not effectively remove weeds that were older than a few days (i.e., outside the seedling stage), it could not operate for longer than a few hours, it moved extremely slowly (especially when it was muddy, which was 75+% of the time), and it had so many problems and breakdowns that it required two highly paid people with engineering degrees to monitor and troubleshoot it at all times.
Assuming good weather and no overgrown weeds, this car-sized robot and its two handlers could achieve in 3–4 days what a single minimum-wage employee equipped with a hoe could achieve in 3–4 hours.
That robot was a major factor in my decision to quit. It was one of many very questionable decisions our new director made (including buying a wildly over-specced drone, which required two people to fly at all times — one to actually fly it, and one to watch for angry eagles that would soar in from miles away to come knock it out of the sky because it was so loud), but the robot really topped the cake.
Edit:
Here are examples of automation that we did use regularly with great success, most of it very tried-and-true by now:
I worked in the R&D department, and automation was very low on our list of concerns (until our new director tried to shove it down our throats). Even though Australia has very high wages, the cost of labor was not our biggest hurdle, nor the biggest hurdle faced by our customers (major commercial farmers worldwide) — not by a long shot. I'd say that about 95% of our efforts were biology-focused: combating disease, improving yield and quality, increasing pollination, expanding climate tolerance, etc.
Really fascinating stuff! Thanks a lot for sharing. I'm clearly in way over my head but it's cool to read about. It sounds like farms and warehouses have things in common, as well as vast differences. In common are that automation is most effective and foolproof when it comes to controlling variables of an enclosed space. And manned machinery is widely deployed to great effect. Difference is the warehouse is a cleanroom, vastly different from the outdoor environment, so mobile robots can actually work well in warehouses. Whereas locomotion is the achilles heel of the farm robot, directly contributing to breakdowns and jams.
State of the art is automated steering on big tractors.
I'd be surprised if there's any small-scale automation outside of education. Productivity improvements in farming come from larger plots of land, and people who garden at home generally do it because they enjoy the process of gardening with their hands.
Well, not entirely by hand. Rototillers are common.