Trundling across uneven soil, a machine about the size of a small dog picks its way through rows of young plants. As it es each seedling, it glowers at it with an unblinking “eye”, gathering information about its health and maturity, and looking for signs of disease. Equipped with a camera and other sensors to assess attributes like soil quality, this diminutive, insect-shaped robot is Tom – a prototype autonomous farm assistant. It is a glimpse of what the future of farming could look like.

Barriers to new farmers entering the profession, environmental degradation and a growing world population are putting increasing pressure on agriculture to do more with less. But farmers and researchers are working on increasingly innovative high-tech solutions to find ways to produce more food – enough to feed 10 billion by 2050 – while minimising environmental impact.

Robots like Tom are hoped to be a part of the solution. The diminutive buggy is one of three robots being developed by UK-based agricultural technology start-up Small Robot Company. While Tom monitors how the crops are growing, another robot the size of a small car, called Dick, will move through the field to target any weeds that might spring up. The third robot ⎼ called Harry, of course ⎼ will be responsible for planting the seeds in the first place.

“We are aiming for a new era of precision farming where crops are not cared for at the field level, but at the individual plant level,” says Sam Watson Jones, one of the founders of the company. “It is something that is not possible when you have humans doing the work or with big farming machinery that are currently used.”

The company currently has two Tom robots working in the fields of 20 farmers that have signed up to help trial the technology. They have a prototype for Harry and are in the process of building Dick. In the next year they hope to have developed a series of modular toolbars that will be fitted to Dick and Harry so they can be adapted to different tasks in the fields. The firm is also testing a prototype of a central control system that will gather data from the robots, called Wilma. Watson Jones began looking at how robots could help to change farming after noticing the rising costs of machinery on his own family farm in Shropshire. Looking back over 25 years, he noticed that the crop yields from the family land had not changed. Their profits instead were subject to fluctuating crop prices. “At the same time the cost of machinery was going up every year,” he says. “I began talking about it to other farmers and everyone had noticed the same trend. It became obvious to me that the way we were operating had to change if we were not going to eventually go out of business.”

His hopes for a transformation of farming are not unfounded. Through the centuries, innovations and new technologies have not only had a dramatic impact on how crops are grown and the amount of food we can produce, they have also moulded society.

But while human judgement might always be needed to some degree, autonomous robots could take a lot of the hard, repetitive work out of farming.

Engineers at Harper Adams University, in Shropshire, UK, for example, have managed to use autonomous vehicles to sow, grow and harvest a field of wheat without humans ever having to touch the plants. Back-breaking work like weeding can also be done entirely by robot. With the help of artificial intelligence and onboard cameras, some can pull the weeds from the soil as they move along rows of vegetables.

Switzerland-based ecoRobotix’s solar-powered weeding bot precisely targets unwanted plants with herbicide, reducing excess runoff that can pollute ecosystems. Similarly, Blue River Technology in California has developed a machine that can identify weeds among rows of cotton plants, selectively spraying them with herbicide. It is also developing a system that will simultaneously spray fertiliser onto the cotton, transforming traditional mass crop spraying into a targeted process that reduces the amount of chemicals used in fields.

Machine vision is also helping farmers to detect the early signs of disease in their crops. Ernest Mwebaze and his team at Makerere University in Kampala, Uganda, are using AI to detect four key diseases that affect cassava plants on farms in East Africa.

“The big problem is differentiating the four diseases as they need different responses,” he says. Their system identifies the disease and its severity from simple camera phone photographs through an app. “We are trying to give farmers access to expert knowledge in their pocket,” Mwebaze adds. In the US, scientists at Cornell University are also using robots armed with laser scanners, multi-spectrum cameras and AI to help farmers prune leaves and fruit from vines to help improve grapes used in wine making.

Harvesting too is about to change. Dutch agricultural machinery manufacturer Cerescon has created a robot that picks white asparagus, keeping the vegetable out of sunlight that can damage the crop as it does so

Engineers at Queensland University of Technology in Brisbane, Australia, have developed a sweet pepper-picking robot called Harvey, while the Orchard Machinery Corporation in Yuba City, California, have created a robot that can shake almonds and apples out of trees.

The last agricultural revolution, driven by innovations like the combine harvester, welcomed an era of intensive, industrialised farming. The next revolution will likely be one that delivers precision and lessens the impact agriculture has on the environment.

According to the International Food Policy Research Institute, using precision farming techniques and new technologies could help to increase crop yields by up to 67%. Goldman Sachs estimates 15-20% of yield is currently lost due to imprecise fertiliser application while 40% of fields are over-fertilised.

Applying fertiliser, herbicides and pesticides in more targeted ways could help to reduce costs for farmers but also ensure plants get the right amount of each treatment. Precision farming like this may also necessitate a switch from giant farm machines to smaller autonomous equipment. This too could bring dividends.

“Using large, heavy tractors that drag ploughs through the earth can have a negative impact on soil health,” says Sam Watson Jones of Small Robot Company. “It kills worms, damages the fungi and microbes that help plants grow and releases a lot of nitrous oxide in to the atmosphere.”

By using the power of the digital age to handle large amounts of data, crops can also start to be dealt with as individual plants rather than on a field-wide basis. Assessing their health and needs plant-by-plant rather than as an average could mean different crops can be planted side-byside on the same patch of land, and then harvested separately when they are ready

“So much of the countryside looks the way it does because of the machinery we use for growing food,” adds Watson Jones. “If we begin to grow crops in different ways then we could change how the countryside is currently designed to a form that is better for us and for the environment.”

The plough. The tractor. The combine harvester. These innovations have allowed our rapidly expanding species to feed itself and thrive. By sending robots, drones and algorithms out into the fields, farming now sits on the cusp of another technological milestone.

Welcome to the smart farming revolution.

Image credits: Getty Images, Small Robot Company, Stara

Graphics sources: Our World in Data, Goldman Sachs Group

--

This article is part of a new multimedia series Follow the Food by BBC Future and BBC World News. Follow the Food investigates how agriculture is responding to the profound challenges of climate change, environmental degradation and a rapidly growing global population

Our food supply chains are increasingly globalised, with crops grown on one continent to be consumed on another. The challenges to farming also span the world.