Are small swarm robots the future of agriculture? Harry Henderson, knowledge exchange manager, argues there are some fundamental questions that need answering before our big machines become a thing of the past.
Speak to people today and the latest aid to human life will be the robot. It is fair to say you can already buy a robot to mow your lawn, vacuum your house and, closer to agriculture, milk your cows.
So it’s evolution rather than revolution, perhaps. But field-grown crops are largely the preserve of manned diesel-powered machines, planting the crop, applying crop protection and harvesting the saleable produce. An ideal target for robotics to take over, applying an as yet unheard-of level of attention to detail, going from the working width of the sprayer to analysing each and every plant, and taking appropriate action on that very plant.
But hang on, it’s not quite that simple. Standing in the way of this revolution are many human thoughts and feelings such as simply remaining in business. Let’s not forget that food is grown very cheaply right now, and the consumer is not likely to pay more because a robot needs funding.
Earlier this year, Agri-Tech East’s monthly pollinator meeting focused on swarm robotics with a top agenda of speakers, who not only championed, but also questioned, the use of swarm robotics in agriculture. A refreshing meeting where the potential was tempered by the challenges.
Four challenges to the uptake of swarm robotics:
First up to speak was Dr. Alan Millard, lecturer in robotics, University of Plymouth. Dr Millard reasoned that swarm robots would communicate with each other via Bluetooth or Wi-Fi to identify and treat many plants at a time, while monitoring the crop for other issues, ripeness and harvest prediction. Rather scarily, he argued that swarm robotics are inherently robust because they employ ‘decentralised control,’ meaning that no one robot is in control; if one fails, the rest carry on. They can also monitor each other’s performance, meaning that a poor performer can be removed by the other robots for repairs.
Dr Millard cited four challenges to the uptake of swarm robotics in agriculture.
- In field environments, is Bluetooth good enough, and will 5G networks be as good as we need?
- Capacity to make ‘live’ decisions. Will robots be able to sense crop problems and identify and apply a solution as they work across a field? All this, of course within parameters set by the grower.
- Public acceptance. Is it right that robots are deciding the standard of food quality that the consumer gets to buy? It’s important that the public are on-board from the outset.
- What is the business model? Who owns the technology, maintains and repairs it? Importantly, who has access to the data the robots are collecting in the course of their daily work?
Dr Millard felt that two things are certain. Ground-based technology will win over drone-based tech, as legislation tightens up over aerial operations and limited run times. And modular robots will be able to be adapted to different tasks as the crop demands change.
Closer to today, Dr Millard is working with Cornish growers on autonomous, coordinated harvesting of cauliflowers in collaboration with the Eden Project, Agri-tech Cornwall and Teagle Machinery Ltd.
Clive Blacker, director at Precision Decisions, argues that the current cropping practice is already cheap. The leading trend of reducing labour on farms has meant that bigger machines are being used. The knock-on effect of this is that attention to detail has reduced. “It’s no use claiming that field monitoring can be down to 5 cm squared resolution when your crop sprayer and fertiliser spreader operate at 24 or even 36 metres,” he said.
A recent survey of 4,500 farmers around the world found that 8–12 hours a week is spent entering data in 99 different types of farm software. The complexity and hunger for data has to go down, which will lead to more autonomous decision-making in live, in-field situations.
Are autonomous robots safe?
Mark Nicholson, senior lecturer in systems safety, University of York, spoke about safety management of autonomous farms. From the very outset of autonomous operation, safety has to be built in. Who is responsible in the event of accident or injury? The owner of the technology? The supplier? Eventually, it will be the manufacturers who send their machines to unknown destinations to be operated by persons unknown. Clearly, much more emphasis will be needed in training operators. Mark is involved in the Assuring Autonomy International Program to ensure safety in all types of autonomous systems. This work will be critical in the uptake of driverless vehicles in the field and road.
A more sustainable solution?
Sam Watson Jones, co-founder of the Small Robot Company shared his vision into his work so far. Small prototype robots named Tom, Dick and Harry handling crop and soil monitoring, feeding and weeding, and planting and drilling, respectively, with decisions being made by Wilma, communicating with all field robots. Sam felt strongly that the current size of machinery causing yield limiting soil compaction, along with the loss of active ingredients through legislation and resistance, was unsustainable. Add to his thinking that as much as 85% of UK farms are not financially viable if not government supported, this adds up to a need for radical change on the UK farm. Small robots would increase yield, lower environmental impact and return profitability to farmers. The global farm machinery industry with huge factories to employ would find itself unable to turn around, allowing smaller companies to flourish in this new world of sustainable food production. Bold words indeed.
There’s no doubt the future is coming. How you fit into this new world will be, well, up to you.
image: Concept image of the Small Robot Company’s Dick digital weeding robot (c) Small Robot Company