Wet weather and COVID-19 brought Hands Free Hectare’s autonomous operational plans to a halt this season. But the team has continued with its research, putting in place a new strategy to get going again later this year.
Summer 2020 should have seen the UK’s pioneering Hands Free autonomous farming project harvesting crops on its expanded 35ha Hands Free Farm. Unfortunately, unprecedented wet weather and COVID-19 have conspired to constrain its work this season.
In autumn 2019 heavy rain flooded the project’s land at Harper Adams University in Shropshire, England. Like other farms in the area, they postponed drilling until conditions improved enough to sow spring crops. Then, just as they were getting set to get started, COVID-19 struck, bringing operations to a halt again.
Text continues underneath image
“It may seem strange that we could not operate autonomously, but we actually require a team of three people. We need a driver in charge and two spotters to ensure operations are carried out safely,” explains Clive Blacker, one of the founders of the project who is also a director of Precision Decisions and head of business development for Map Of Ag.
Indeed, interestingly, he doesn’t see autonomous operations actually reducing staff on farms in the future either. “In many areas, particularly UK arable farms, labour forces are already at the bare minimum. We will always need people to run the system and manage the infrastructure, such as ensuring equipment and inputs are in the right place at the right time.”
So instead of watching an autonomous combine working its way through the crops, the team is, like many others, confined to offices and meeting over Zoom to progress it’s plans for harvest 2021. Out on the farm the land is protected under cover crops – providing another research opportunity for the team.
“The enforced break, however, has allowed us time to think and develop new systems and progress with ‘smart’ technology. It also helps us to prepare for this coming establishment season. In the past three years we have autonomously sown, grown and harvested two crops in a single field. This year the plan was to expand that to a farm scale, with different crops and fields over a total of 35ha,” he says.
Text continues underneath image
Moving to a farm scale introduces new challenges that may at first appear straightforward, but can be actually quite difficult. “For example how do you get to the field and back? For manned machines this is simple, but for us it means autonomously opening and closing gates,” he says.
Ultimately we need to also include ISOBUS control into autonomous machines
Hands Free Farm, working with technology partner FarmscanAg, is also spending this time developing new systems. “We are using some existing technology, but some of what we need doesn’t exist or isn’t able to easily work with or integrate into other systems. Smart technology needs the machine to interpret what its detects and then be able to make the appropriate machine adjustments. Ultimately we need to also include ISOBUS control into autonomous machines,” explains Clive.
Safety is a top priority and one that is made considerably more difficult when operating outside the confines of a single field. There are, for example, 1.5km of public rights of way over which the autonomous vehicles will be travelling.
The team organised a meeting discuss rules and regulations with representatives Health & Safety Executive, Agricultural Engineers Association, Institution of Agricultural Engineers and other interested parties.
“Our operations already comply with the ‘Code of Practice governing trialling automated vehicles technologies in public’. But this is mainly aimed at operating on the road,” he explains. “There is nothing that specifically covers working in fields and off-road so we discussed creating some standards, risk assessments and guidance because, we expect autonomous operations are more likely to begin in the fields rather than on the roads.”
Iseki UK is supporting the Hands Free project, which has converted the tractors to run autonomously to carry out all the field work. (Text continues underneath video)
At the same time it has been investigating how to ensure autonomous operations are also ‘smart’. “If you step in front of an auto-steered machine following a route, it will run you over,” says Clive. “A smart machine will detect your presence and stop, but the next question is how does it know it is safe to continue and, moreover, is it able to start itself again?”
On a combine harvester, he continues, an operator is likely to see a green patch or the crop bunching up on the header and will slow down. If it does become blocked, they will back up, reverse the auger, clear the blockage and then ease the crop slowly back into the machine.
“Can you imagine how many different sensors and activators are required for just that simple sequence? It’s not only the physical operation it all requires code and logic to make it actually happen, then check that it has happened,” he continues.
In the past few months the team has been working on preparing its ‘new’ combine – a tracked Claas Crop Tiger 40. Designed originally to harvest rice, it’s small enough to suit the hands-free harvest, but is also about as manually-controlled as combines come.
Its size and tracks play an important part of the research, which is also looking at how technology can be used to reduce compaction and improve soil structure.
“Although the auto-steering is already done, there’s quite a lot to do to enable the combine to operate autonomously,. It’s not just the simple controls like forward/back, header raise/lower, reel and other settings,” Clive explains. “Smart control will be the next step.”
Text continues underneath image
“I’ve also noticed over the years that as machines get bigger, the level of precision reduces,” comments Clive. “When I first started yield mapping with my Claas Mega in 1998 it had a 4.5m wide header. The current 9.0m wide cutterbar on a Lexion 760 is taking a larger swath, plus the combine is travelling faster and as a result I get 60% less yield data.”
“This made me think about the resolution of the data and what is now appropriate? Currently the trend is to vary inputs in ‘zones’ of one soil type – which are often virtual hedges following previous field layouts. Now, however, the highest practical resolution for applications is 50cm from a single nozzle on a sprayer today.
The accuracy is falling as machines get wider and faster
“While we can apply plant protection products at a high resolution, we can’t measure the effect on yield maps as combines get faster and wider. We sow wheat at about 600 ears/m². So a 10m wide cutterbar, travelling at 1m/sec is harvesting 6,000 plants every second. But what if the yieldmeter takes a reading every five metres? That is now 30,000 plants – the accuracy is falling as machines get wider and faster,” he explains.
But, he adds, using the Claas Tiger’s 2m wide header and taking a reading every 0.5m delivers a map with a yield point for every square metre. “This makes it easier to spot temporal differences, which are averaged out by wider machines and, by using a 2m wide drill and individual nozzle control, we can target treatments down to the square metre,” he adds.