Agricultural robot

A single strawberry farm in the UK can span 200 hectares and need thousands of pickers working twelve-hour days across a narrow six-week harvest window. Finding that labour is increasingly expensive and difficult. Missing the window — because a harvest came early or workers did not show — can mean a year's income rotting in the field.

This pressure, multiplied across every crop and every country, is what is making agricultural robots one of the fastest-growing areas in all of robotics.

What agribots actually do

Farming has dozens of distinct tasks, and different robots handle each:

• Planting and seeding — autonomous tractors or small wheeled robots place seeds at precise depths and spacings, often with centimetre-level GPS accuracy.
• Crop monitoring — drones and ground robots carry multispectral cameras that detect nutrient deficiency, disease, and water stress plant by plant, before the damage is visible to the human eye.
• Weeding — robots identify weeds among crops using computer vision and either spray a micro-dose of herbicide on the weed only (vastly reducing chemical use) or physically remove it. Carbon Robotics' LaserWeeder uses a laser to kill weeds without touching the soil.
• Harvesting — the hardest task. Fruit picking requires a gentle touch, three-dimensional vision to locate ripe fruit among leaves, and speed. Agrobot, Tortuga AgTech, and several others have commercial strawberry and tomato harvesters, though full reliability at commercial speed remains challenging.

A real example

Tom, Dick, and Harry — three robots built by the UK's Small Robot Company — work as a team. Tom is a lightweight monitoring robot that scouts the field and maps every single plant. Dick uses that map to apply micro-doses of herbicide or fertiliser to individual plants rather than broadcasting chemicals across the whole field. Harry, the harvesting robot, is still in development. The system exemplifies a shift from field-level farming (treating everything the same) to per-plant farming (treating each plant as an individual).

Why precision matters

Conventional farming applies water, fertiliser, and pesticide uniformly across entire fields. A robot that responds to individual plant data can reduce pesticide use by 90% in some applications, cut water use significantly, and improve yields by catching problems early. This is the promise of precision agriculture — and robots are the tool that makes it practical at scale.

The barriers are real. Farming environments are unstructured and unpredictable: mud, variable lighting, irregular plant shapes, and extreme weather all make robotics harder than in a controlled factory. Upfront costs remain high. And not every farm is large enough to justify the investment.

The same computer vision that helps a robot distinguish a ripe strawberry from an unripe one is trained on millions of labelled images — meaning farmers worldwide are, unknowingly, teaching robots to replace them.