Robotic farming has been around for a while, mainly in greenhouses and specifically in horticulture. For larger outside fields there are robots which can perform a single or several tasks. But they are not very fast.
My (solar) robotic farm does most tasks, independently, continuously and efficiently.
And above all, quite fast. It is precision farming in the extreme.
It can be done by a very simple design which separates it from current robotic systems.
-the big disadvantage of wheeled or legged robots is that they can only move on the ground in between rows. They can only go from point A to B if those points are connected in a straight line. To go from point A to, say, F (which is located in another row), the robot has to travel a long distance. We won't even talk about mud and different types of terrain that may make such a robot's task difficult.
-my robot can travel from A to F in a straight line (like a bird), thus saving much needed power. And it is never bothered by mud or terrain constraints.
The idea is simple:
1. suspend the robot on a cable (let's call this the robot-cable). The robot can travel on this cable by a simple wheel, electric motor and pulleys.
2. This robot-cable itself is suspended on cables which form a square frame, that spans the 4 corners of the field. The robot-cable as a while can travel on these cables, again by being pulled with pulleys and electric motors.
3. the poles, which hold the four main cables, can move vertically (telescopic poles); and so, as the crop grows, the robot can always hover above it.
Thus, like a scanner, the robot can move to any point on the field very quickly, and, in this system, even vertically. That is, we can control movement in /// 3 dimensions ///.
The entire system is electric and solar powered.
Digital:
-The robot accurately "tags" each individual plant, and so exactly knows where to analyse or perform a task (this tagging occurs during the planting, also performed by the robot.)
-The robot therefor does not need GPS, as it calculates spacing itself (it knows it planted seed 1 at a particular point; seed 2 is located to the North, 40 centimeters further in the line). So the robot makes an entire grid.
-In a sense, this is a "digital" grid, in that the robot can pinpoint an individual plant like a dot on the grid
Many small implements can now be attached to the core robot, to perform basic tasks:
-vision sensors (infrared camera, etc...), used to track growth, quality and threats (leaf color, insects, etc...)
-a seeder: a small tube that can dig into the soil and deposit a single seed; this allows for no-till farming
-a fertiliser: similar small tube that can dig and spray a drop of fertiliser; for example, after 2 weeks, it can deposit a few grains of urea, at a distance of 5cm from the seed-point; after 6 weeks it can deposit again, but now 10cm from the seed point, and deeper in the soil - this extreme precision will save a lot on fertiliser
-a weed removal device: a simple digger and scissor of sorts that can dig up or cut weeds (there is already a weeding robot out there)
-an insect suction pump; the robot can, via pattern recognition, detect bad insects, and suck them up with its insect vacuum cleaner; this bad biomass can then be collected and used either as fertiliser or energy after conversion (anaerobic digestion or composting)
Depending on the task, the robot will itself go fetch the needed implement. It will travel to the main tools hub (located in a corner of the farm), where it can pick up the tube, pump, or whatever.
The robot can work day and night. Power can come from either a battery carried by the robot itself, or from a battery located outside the field (the robot would then carry a cable connected to this battery). I designed one with the robot carrying its own solar panels and batteries.
In short, the idea offers a robotic farming platform that can be adapted for any crop and that uses no fossil fuels.
All the components and software to make this scanner are available in electronics and hardware stores. Mother boards and functions for servos can be designed with programs like Arduino.
I made a small sketch of it (see link).
PS: the only task that may be difficult is harvesting. But perhaps we can design a tool for that too. However, the biomass at harvest is heavy, and the tiny robot cannot carry this load in time, to prevent the grain or other product from rotting.