Shi Yan’s approach to organic farming is helping to break the country’s “addiction to pesticides”.
For a change, Kevin Sanders has decided to let someone, or more accurately, something else count the apples in his orchard. This isn’t the first time his idyllic farm down in Australia’s Yarra Valley has played host to robots and their handlers, so Sanders knows what to expect.
Moving soundlessly down the corridors between trees, an electric robot will scan each plant, identifying individual fruit and flowers. An algorithm is then used to classify and count the apples in each image and provide a yield estimation, a critical figure for farmers that will inform Sanders’ plans to manage his orchards and the harvest.
A fourth generation farmer, Sanders and his brothers have an interest in innovation that has created an unconventional operation.
The winner of a Hugh McKay Future Farming Award in 2010, the Sanders’ farm is a particularly good fit for robotic field trials because of their unusual orchard design.
The trees here grow flat on a V-shaped trellis making them 2D rather than 3D, where the branches are not allowed to grow in all directions. They are short and typically only a metre wide. “These aren’t the dimensions people grew up with,” says Sanders, adding however that it can still take hours and hours to count blossoms and fruit and carry out essential activities such as thinning out the branches.
All this time adds up.
“It’s not uncommon to end up paying anything between A$1,000 to A$7,000 (US$737-$5,168) a hectare in labour bills,” Sanders tells Al Jazeera, explaining that in Australia hired labour typically accounts for the largest percentage of cash costs, making robotic solutions appealing, particularly as local farmers’ scramble to remain competitive in international markets.
And there are other concerns. According to data from the Australian Bureau of Statistics, Australia’s farmers tend to work well beyond their retirement age. In 2011, almost a quarter were over 65 years old, compared with only 3 percent in other occupations.
The figures also reveal that the growth of large-scale farming operations and increasing urbanisation, with more young people moving to cities to live and work, has had a detrimental impact on farmer numbers.
An analysis of the 30 years leading up to 2011 revealed that the number of farmers declined by 106,200 (40 percent) – that’s an average of 294 fewer farmers each month.
Salah Sukkarieh, director of research and innovation at the Australian Centre for Field Robotics (ACFR) at the University of Sidney, thinks that the answer might lie in increasingly sophisticated agricultural robots.
The past decade has seen widespread adoption of digital technologies such as sensors on Australian farms. But the real game-changer has been the leaps forward in computing power that have enabled farmers to receive and act upon this data in real time.
“It’s an easy extension from there for robotics to kick in, because if you can sense in real time, then you can make decisions in real time and you can act in real time,” says Sukkarieh.
ACFR’s first forays in agricultural robotics began with the tree crop industries, where farmers wanted to switch from making decisions on a macro scale to a micro scale.
“They were interested in knowing how individual trees were performing, which they had never been able to do before,” he says.
The robotic system Sukkarieh and his team built for its Shrimp robot could collect LiDAR – light detecting and ranging – visual, thermal and hyperspectral data as well as track soil conductivity and natural gamma, “demonstrating that there are many ways to view the humble tree”.
In 2016, the ACFR’s SwagBot had its first field trial. Whirring madly, the robot was able to pull itself out of ditches, roll over branches and traverse stretches of water.
Set to work herding cattle, SwagBot got them all moving in the right direction. Sukkarieh imagines that SwagBot, which can also be used to autonomously tow heavy loads such as feed into the field, could help people manage sprawling cattle farms even in very remote locations.
Sukkarieh says the most natural next step will be to teach the robots to identify sick or injured animals based on variables such as their body temperature. Because pasture quality is key for high-quality milk, farmers would welcome soil samples and other measurements from grazing areas.
ACFR also has drones, and it is possible to coordinate between ground and airborne robots. Currently, ACFR has been deploying its aerial technology to detect weed infestations on both large and small farms.
Sustainability has been an important focus for the ACFR team. Their RIPPA robot has been put to work on several commercial vegetable farms across Australia, where it trundles along rows of vegetables, conducting autonomous, real-time soil sampling, weed identification and de-weeding, as well as dispensing water, pesticides and fertilisers according to the needs of individual plants.
With RIPPA able to use satellite-based corrections to operate within 4cm precision, this has radical implications for irrigation management and agro-chemical use on the farm.
Earlier in the year, the robot had its first endurance trial during which it completed nearly 22 hours of continuous operation, roving autonomously up and down spinach crop rows, using only battery and solar power. Its batteries finally died in the early of the morning, but when the sun rose, the machine recharged and was back in operation.
Anthony Kachenko, the R&D lead at Horticulture Innovation Australia, saw what RIPPA could do first hand. He believes Australian farmers who are forewarned can be forearmed.
“Being able to detect disease at the onset – or even predict conditions where there is more risk of disease – will help with disease management and potential savings in chemical use,” says Kachenko.
He adds that data from agricultural robots have already helped farmers identify hot spots where further attention is needed. “This is just the beginning, as the next step is providing the remedial action to help the farmer farm smarter.”
However, there is still some way to go before such robots become common on farms.
“We haven’t reached the dexterity of what a human hand can achieve,” Sukkarieh tells Al Jazeera, explaining that one of the big concerns remains cost. “In agriculture a farmer typically doesn’t have a lot of money, the margins are very, very small.”
Rohan Rainbow, managing director at Crop Protection Australia, says the new technologies might actually inspire a new generation of young farmers, who are enthusiastic about technological innovation, to return to the field. But the effect on labour and how the technology is implemented is still to be seen.
“As you become more automated, the way people interact in the field and with equipment changes,” he says, “and there seems to be a need to develop social license, if you will, for commercialisation of those technologies.”
In fact, the industry’s big challenge might not be the technology itself but satisfying regulatory requirements of government safety and giving confidence to the insurance industry that the machines can be operated safely, says Rainbow.
There is also the matter of enabling real-time communication and exchange of huge data packets through communication channels between farmers and the robots and between the machines themselves – many farms are so isolated that access to bandwidth is an everyday challenge.
Speaking at the Falling Walls Conference in Berlin, Sukkarieh told his audience that his team’s goal was to have farmers be able to run all these systems directly from their smartphones.
As the technology is widely used it will also become increasingly affordable, putting it within the reach of farmer cooperatives and government bodies even in developing countries.
“Think about 3D printing and how that is going to make these components much cheaper to manufacture,” he said.
Sukkarieh has also started to think about what 2050 will look like from a robotics perspective. The requirement for sustainable operations, labour shortages, 24/7 precision agriculture, minimising costs and minimising chemical use and energy will all shape this burgeoning field. Not too far in the future, Sukkarieh imagines the possibility of semi-automated or even fully autonomous farms.
Grains, vegetables and fruits could arrive in supermarkets without a single human ever having touched them. “The question is how do you feel about that?” asks Sukkarieh.