Open the door to this understated cabin and there’s an incongruous sight: monitors stacked atop computers, keyboards, graphs, charts, and a view onto a very nondescript field, next to a university’s rugby pitch, where supporters are going wild because someone has just scored a try.
I am not far from the village of Edgmond in Shropshire, deep in rural England, on the edge of Harper Adams University, which specialises in education in the agricultural and rural sectors.
So what’s all this – the future of farming? Is this the equipment? Where are the farm hands? The answer is that there are none. Because this field is actually very special: it will soon yield a crop that will be the first grown without interference from a human. Every process – from preparing, fertilising and drilling the soil to monitoring, maintaining and, ultimately, harvesting the crop – is autonomous and performed by specially customised machines: self-driving tractors, robot combine harvesters and drones.
The UN expects the world population to reach 8.5 billion by 2030 and 9.7 billion by 2050. To sustain this growth and prevent young people deserting farming, many believe we need another agricultural revolution. While self-driving tractors have been available for several years now, it’s only recently that, particularly in the US, economic factors have been forcing increasing numbers of farmers to look to autonomous and precision technologies. To meet demand, R&D is gathering pace. The future is all about 24/7 fleets of driverless agricultural machines and drones equipped with sensors and infrared cameras to monitor soil variation, detect pests and eradicate disease.
But bringing these state-of-the-art technologies together, to work over a complete arable crop cycle, has never been attempted before. That’s why the world is watching what’s happening on a parcel of land next to a British university’s rugby pitch. The Hands-Free Hectare is undoubtedly an ambitious plan: to sow, grow and harvest a crop of spring barley using only readily available machinery, autonomous control systems and open-source technology, with the untried systems undergoing rigorous lab and field testing beforehand. If it happens, it will be a world first.
To find out more, I track down Kit Franklin, a lecturer at Harper Adams and one of the three engineers behind the concept.
It proves to be no easy task. The team, which also includes roboticist and Harper Adams teaching assistant Jonathan Gill and Martin Abell of high-tech farming specialist Precision Decisions, has been working flat out to prepare the equipment for seed drilling. But, just like bog-standard farming, the process is weather-dependent and it’s been too windy to spray on glyphosate, the herbicide that kills off grass and weeds. They’ve had to wait and worry.
“It’s been a mixture of stress, excitement and anticipation,” Franklin tells me, when I finally manage to arrange a chat. “If we don’t get our crop in the ground, the project fails – so it’s all hands to the pump to make sure we’re doing what we told the world we’d do! While I’m speaking to you, the chaps are out in the field with the drill, setting it all up.”
To make the tractor drive autonomously, ‘waypoints’ are plotted onto a chart and uploaded onto the control system, housed in the cabin – the centre of operations. These ‘dots’ tell the tractor where to drive and where to turn. “It does dot to dot, basically,” says Franklin.
The drill attaches to the back of the tractor via a three-point linkage. When it gets to the desired location, a GPS receiver on the tractor sends a signal to the linkage to raise or lower the drill at the right place. The system uses Real Time Kinematic satellite navigation, the most precise available, with accuracy down to 20mm.
At the start, the drill is dropped into the ground, the tractor performs a run and then, in response to the signal, the linkage lifts the drill. The tractor then turns around, with the drill suspended in the air, and steers towards the next waypoint where it needs to drill again. A signal is then sent to lower the linkage and the process is repeated. When the field is harvested, the head on the combine harvester will perform in a similar fashion. “It all sounds very simple but there’s a lot needed to make it happen,” says Franklin.
Harvesting, particularly selective harvesting – where only ripe and ready produce is picked – is especially problematic for machines. So researchers, established companies and start-ups are working hard to develop, pilot and launch innovative systems to tackle this and a wide variety of other common agricultural tasks. The EU has funded at least six projects around robotic farming, including cloud-based MARS (Mobile Agricultural Robot Swarms) and RHEA, a scheme to develop a fleet of tractors and aerial robots for targeted herbicide application.
“For the last 20 years, people have been talking about farming with robots,” says Franklin. “But nobody has yet put a whole system together – the entire growing cycle – all done with robots in one project. We wanted to stop talking about robot farming and do robot farming. We wanted to show that it’s possible.”
Franklin’s family background is in farming. He credits this for his love of machinery and early desire to design and develop tractors. While pursuing a degree in agricultural engineering at Harper Adams, he became entranced by the possibilities of robotic machines and now wants to inspire future generations and dispel the flat-cap-wearing image of farmers. “It’s high-tech and there are exciting jobs to be had,” he says.
But won’t there be less, not more, work if systems are automated, I wonder.
“We’ve taken the farmer out of the field but you’re always going to have to understand your soil and your crop, and I don’t think there will ever be a replacement for having to get a bit muddy now and then,” he laughs. “The technology’s just going to make it easier and make you a more efficient farmer.”
In fact, this is the real USP in robotic farming: taken together with precision agricultural techniques, it has the potential to revolutionise the way farmers address the perennial challenges of crop production, such as controlling problem weeds with automated mapping and targeted herbicide application.
Milestone in robotic farming
More accurate yield mapping – one of the foremost techniques in precision agriculture – is also possible with the smaller machines common to autonomous farming. Yield monitors use sensors attached to combine harvesters to measure the amount of crop collected. This data is located, recorded and analysed to improve productivity. But accuracy is limited to the width of the combine’s header. In smaller machines, blocks of data are accurate to within 2m, compared to 12m for some commercial vehicles.
One field removed from the Hands-Free Hectare is a commercial farm where a full-size tractor trundles along, its driver just visible at the helm. In comparison, the 40hp man-less Iseki tractor, selected by Franklin and his team, seems lilliputian. David and Goliath spring to mind.
These smaller, lighter machines have another enormous benefit: they don’t damage the soil to the degree that modern full-size tractors do. Over the years, tractors have grown larger to counter the tight working window that farmers have to navigate with a small labour force, but soil compaction has been a side effect. Healthier soil and more precise, targeted inputs lead to higher yields. And, because the small machines are autonomous, one person can take charge of three or four.
Under its head of engineering, Simon Blackmore, Harper Adams has pioneered the development of a fleet of smaller, lighter, cheaper agricultural machines. Multiple semi-commercial prototypes have now been demonstrated by agricultural machinery companies. Blackmore, Franklin’s former boss and teacher, is proud of what his protégé is striving to do. “The Hands-Free Hectare is an important milestone in the development of robotic agriculture,” he says. “It shows, for the first time, that crops can be successfully grown with small, smart machines.”
But plenty of other research questions remain, says Clive Blacker, managing director of key project partner company Precision Decisions, and a Harper Adams alumnus. “What we really didn’t anticipate is how much we didn’t know and how much we couldn’t actually do at this stage,” he says.“There are benefits which we aren’t able to quantify in this project, such as the positive effects on the ecosystem and the soil. We can show that smaller machines are effective but the million-dollar question is how many do you need and is it cost effective?”
While very enthusiastic about the project, Blacker is also prosaic about the commercial realities. “While there’s a huge amount of interest in autonomous vehicles, applications in agriculture are complex because of the need for rate control, for instance, and making sure the right things are going in the right places,” he says. “I do think robotics, on a small scale, will potentially change the way we farm but I don’t think it will be overnight.”
The next step is to work out commercial uses. However, he adds, “in high-value crops, where there’s demand on labour, or in controlled environments, these robots could be in use fairly quickly”.
In the US it’s already happening – a combination of increasing costs and a shrinking labour pool is nudging farmers to take a leap into precision agriculture. A report by Tractica, a Colorado research firm, predicts that shipments of agricultural robots will increase from 32,000 units in 2016 to 594,000 by 2024.
David Leaver, president of the British Institute of Agricultural Consultants, is convinced that “sustainable intensification of crop and livestock production systems will require a high-level, precision farming approach, and agricultural engineering R&D will be pivotal to this achievement”. Farming could even become a 24-hour process, uninterrupted by darkness or the need for sleep. Although the technology is readily available, without the scope to use photovoltaics, improvements in battery life are a prerequisite.
But development requires money, determination and a great deal of hard work. Funding for the Hands-Free Hectare is being provided through Innovate UK’s Satellites and Agri-Food Competition and various businesses have lent equipment and expertise. Even though no one goes in the field, there are a lot of hands at work behind the scenes of the project. And, with all that investment and worldwide interest, it has a lot to live up to; the pressure is on.
The next big task is to adapt the commercial-grade combine harvester to function as an autonomous machine. The team plans to harvest by early September and Kit Franklin is hoping for a six-tonne yield. “That’s not by any means the best yield in the world but, this time round, it’s about proving the technology. Anything we get at the end is a bonus,” he says.
“I imagine it’ll probably be a couple of months after harvest but, if we can get some beer made, I think we may have to have a party. Although that’s a long way away, to say the least!”
Thorwald – the ‘Brexit bot’
Within the UK’s farming and food production industry, the fear of post-Brexit labour shortages is palpable. The fresh produce sector is particularly reliant on seasonal, migrant workers, already in short supply. To plug the gap, one of the country’s largest produce companies is behind a major initiative to develop a multi-functional farmbot, Thorwald.
Adept at tasks such as delivering trays of strawberry plants and providing a night patrol service to kill off crop-spoiling mildew, Thorwald is the creation of Lincoln University robotics professor Pal From. The plan is to use the unnamed company’s funding to employ 30 scientists and create a fleet of Thorwalds to man the UK’s farms.
But selectively harvesting fresh produce is difficult for robots. Vision systems are needed to identify ripe fruit, together with bespoke, agile grippers so it remains undamaged. Commercial innovation is gaining pace: in Japan, a strawberry-harvesting robot is now roughly equal in cost to a human labourer, provided that several farms share the investment.
Hands-free hectare timeline
Project started with testing of the automation system on an electric all-terrain vehicle.
From November onwards:
Autonomous system incorporated onto the 40hp Iseki tractor used for drilling and spraying the crop.
Hectare sprayed with glyphosate herbicide to prepare soil for crop.
25 April (Big day):
Hectare drilled autonomously; crop established.
After 25 April:
Team focusing on adaptation of combine harvester to work with autonomous system.
“Crop-walking” via drone. Data gathered enables application of fertiliser, herbicide and fungicide as needed, via autonomous sprayer.
End August/early September:
Harvested spring barley malted, brewed and turned into beer. Party?