From the pages of Produce Business UK

A prototype robotic raspberry harvester developed by Fieldwork Robotics at theUniversity of Plymouth has the potential to be reconfigured to pick a variety of fresh fruits and vegetables, and repurposed for other field tasks, as well used for gathering vital data to improve crop management. PBUK speaks with a number of those involved to discover how the technology could support the UK horticulture industry.

By firstly tackling the challenge of harvesting raspberries – the most delicate of soft fruits – Dr. Martin Stoelen, a lecturer in robotics at the University of Plymouth, is confident his prototype robotic arm can be adapted to harvest other fruits and vegetables.

“We’re trying to go for raspberries, which is one of the more difficult fruits, because we feel we have the robotic arm technology to operate in a more complex environment,” Dr. Stoelen tells PBUK. “If we can crack raspberries, we can do something with other, easier fruits.”

Already, Dr. Stoelen and his research team at the University of Plymouth are developing robotic arms for the selective harvesting of raspberries, cauliflowers and tomatoes; working with Hall Hunter Partnership in Berkshire, Riviera Produce in Cornwall and the Sunqiao and Shanghai Jiaotong University in China respectively.

“We’re trying to avoid developing something that can only be used for one crop, and that is limited by the season of that crop in the UK,” Dr. Stoelen explains.

“The three crops we’re looking at right now are fairly widespread. If we can find synergies, we can probably customize the technology for other fruits and vegetables, depending on funding.”

Fieldwork’s raspberry robot has piqued the interest of David Green, chief operating officer at Hall Hunter Partnership (HHP), which produces 14,000 tonnes of raspberries, strawberries, blackberries and blueberries for Waitrose, Marks & Spencer and Tesco, among others.

As of September [2018], HHP will begin field-testing the technology. The goal is to develop a commercial prototype that will support HHP’s desire to expand production. 

“I’m confident that if Fieldwork can get the raspberry robot to work, they’ll be able to quickly diversify that technology to other fruit,” Green tells PBUK.  

“The state of berry occlusion in a raspberry crop and the sensitivity of the berry are all much more challenging than in a table-top strawberry environment. If they can crack raspberries, I think strawberries will become a lower technical challenge.”

David Simmons, managing director of brassica grower Riviera Produce in Cornwall, is also assisting Dr. Stoelen in the development of a prototype robotic cauliflower harvester under The AutomatedBrassica harvesting in Cornwall (ABC) project. He is equally positive about the technology’s broad potential.

“There is merit from one robot hopefully leading to another,” Simmons says. “If they can crack cauliflower, that robot could harvest cabbages and be adapted from that. They are quite confident they can develop something for courgettes.”

A competitive field

Of course, with interest high among produce growers, many companies are prototyping other promising robotic harvesters. 

Examples include the strawberry-picking robot arms from Octinion in Belgium and start-up firm Dogtooth in the UK, as well as the tractor-like strawberry harvesters, such as the Agrobot which is being trialed by Driscoll’s in the USA, and a similar machine from Harvest Croo in Florida.

Riviera Produce is also trialing a robotic broccoli harvester from KMS Projects.

“This broccoli harvester seems to be a bit further on,” Simmons comments. “It needs some tweaking to get it to do the job. But broccoli is completely different to cauliflower – it’s easier to see the head as it’s open, whereas with cauliflower the curd is wrapped up in the leaves.” 

Having worked in robotics for over a decade, Dr. Stoelen accepts Fieldwork is not the only player.

“This is active research in several UK universities and abroad,” he acknowledges. “There are big players working on selective strawberry harvesting in the UK, Spain and the USA. 

“There are also many companies working on sweet peppers, cucumbers and lettuces. But in terms of a system that is commercially available on the market, there are virtually none.” 

Indeed, although the race is on, it will not be easy, notes Dr. Ed Moorhouse, director of Agri-Food Solutions – a technical and innovation consultancy for the food and farming industry.

“This is obviously a high priority for the UK industry with the imminent restrictions on labor availability, and it is not surprising that growers, such as HHP, are making direct investments in this area,” he comments.

“Finding a cost-effective system that can meet retailer standards and navigate around current crop structures in irregular tunnel environments is a major challenge. I fear that there will be a significant reduction in the UK acreage with replacement by imports before robotic picking is ready for full commercial release.”

Green from HHP, who has seen various prototype robot pickers from other developers, agrees the competition is strong. Nonetheless, he remains highly positive about the technology presented to him by Fieldwork.

“I think Fieldwork has the potential to be a game-changer,” he claims. “I was surprised when I was approached about a prototype for raspberry harvesting. Raspberries are a much softer berry; it’s a much more technically challenging process. But seeing is believing.”

The attraction

On witnessing the laboratory prototype, Green was won over by the robot’s delicate touch and flexibility. 

“A lot of the robotic harvesters I’ve seen have come from the Dutch model of snap picking for strawberries – Fieldwork’s robot was very impressive in the way it travels and moves the raspberries from the plant,” he recalls. 

Dr. Stoelen explains that this is partly what makes his robot stand out. “I do think we have something unique,” he remarks.

“While most robot arms are rigid, the ability to be ‘soft’ is vital in a more variable environment because if the arm hits something it could easily damage that object. We apply the concept of ‘variable stiffness’. Our variable robotic arm can be soft when performing fast movements, and it can be stiff but gentle when handling the fruit.” 

As for the gripper, Dr. Stoelen uses customised finger pads with a mixture of soft materials and “very precise” force sensors. As such, he claims the pads are able to apply just the right amount of pressure to the raspberry without damaging the fruit.

When it comes to avoiding the spread of disease, which is also a problem for human pickers, Dr. Stoelen is looking at enabling disease detection using the sensor package already on the robot. Disinfecting the pads might be an option also, but it remains too early to tell. 

This attention to detail, coupled with Fieldwork’s willingness to work with the industry from the get-go, were other reasons HHP agreed to the trial.

“Fieldwork is very focused on our requirements and what the UK supermarkets need from a quality standard,” Green states.

“They aren’t looking to cut corners to make supermarkets accept a lower quality of fruit. Their aspiration is for the robot to be as good as, if not better, than a human. 

Green believes HHP and Fieldwork could develop “something significant”.

“Six months ago, I wouldn’t have said the tech would be where it is today,” he admits. “I’m confident that in another six to seven months I’ll be amazed at what they’ve achieved.”

Robots: the future?

Both Green and Simmons believe robotic harvesters are the way forward for horticulture. 

“By the mid 2020s, we’ll all be looking at using mechanical harvesting,” Green predicts, noting that the “easier gains” have already been made by most of the berry industry, such as moving to high table tops.

“Certainly, there will come a time in Western economies when it’ll be unavoidable. We won’t be able to maintain customer price expectations with human labour.”

With the support of robots, Green is eager to raise productivity at HHP and expand the business.

“Probably half of our labor is spent on harvesting, so the huge opportunity for me is massively increasing the number of fields we can crop with the same headcount,” he reveals.

Ideally, Simmons from Riviera Produce would like robots to harvest all his crops. For him, addressing skilled labor shortages and standardizing costs are the key challenges to overcome.

“It’s got to be the future,” he comments. “The costs keep going up every year and customers won’t pay any more for the product. We need to somehow reduce costs to remain competitive in the marketplace. Hopefully, robotics can help.” 

These are precisely the benefits that Dr. Stoelen and his team aspire to bring to the marketplace. 

“We can give producers more resiliency to shortages of labor, changes in exchange rates etc.,” he explains. “Growers tell us that harvesting is already expensive, and it’s getting more expensive. For raspberries, manual harvesting can represent up to 40-60 percent of the overall cost.”

Already, Dr. Stoelen claims his robotic raspberry harvester could achieve a lower cost per kilogram than human labor, taking into account downtime.

Whereas human pickers tend to work 40-hour weeks, his current technology could allow for multiple robotic arms to be mounted onto a number of platforms, with each arm picking raspberries simultaneously, and each platform operating day and night for 120-130 hours a week.

Improving crop management

Dr. Stoelen also foresees a broader range of advantages.

“There’s potential to improve the quality of the crop,” he points out. “Robotic solutions have the potential to be more consistent, to reduce risks due to hygiene, etc.”

For Green at HHP, another advantage will come from the 3D cameras and sensors within the robots that can visually log the state of a crop regularly, and on a large scale. 

“Processes like yield forecasting and pest and disease control will be able to be much more targeted because we’ll have a very powerful computer system flagging up the slightest of discrepancies on a day-to-day and week-to-week basis as it passes through a field,” he points out. 

Going commercial

To further advance the technology, the University of Plymouth’s commercialization partner Frontier IP, is actively seeking other trial sites for soft fruits and vegetables.

“We are seeing strong interest from other leading growers in both applications,” claims Rui Andres, portfolio manager at Frontier IP, which holds a 27.5 percent stake in Fieldwork Robotics.

“We are looking for other companies to trial the tech, and are in discussions with a multinational agribusiness about potential applications to harvest vegetables, including cauliflower.”

If successful, Fieldwork plans to offer producers a service-based model on an as-needed basis. 

“The fact that we are designing the technology around a service-based model is fairly unique,” notes Dr. Stoelen. “Maybe a grower lacks 50 harvesters and needs a robot to cover that, or maybe they need a long-term plan. The key point is that a lot of the maintenance and daily operational needs becomes less of a worry for the grower.” 

Over the next year, Dr. Stoelen hopes to develop an alpha prototype of the robotic raspberry harvesting platform. A commercial prototype could follow in two to five years, depending on funding. 

“We’re looking for private investment to push this towards something we can mass produce and put into the market,” says Dr. Stoelen. “You can’t do that with research and innovation funding.”

Green at HHP agrees funding is vital. “If we want to keep having a vibrant and significant horticulture industry in the UK we need to make sure we’re investing and developing these technologies.” 

Simmons at Riviera hopes the government will help. “It’s a small industry, and margins are tight – we don’t have huge amounts of money to invest in the future. Funding is key.” 

Overall, Dr. Stoelen is convinced about the potential for robotics. “The robotic technology developed by us and others could have a significant impact on the agriculture sector ... I think a lot can happen,” he concludes.