Field robots set to make farming faster and simpler

Researchers at The University of Sydney’s Australian Centre for Field Robotics (ACFR) are developing intelligent systems to do everything from detecting and weeding out pests to assessing, tending to and harvesting fruit and veg.

It’s likely such systems will play a key role in the future of Australia’s $46+-billion agricultural industry. And according to Professor Salah Sukkarieh, professor of Robotics and Intelligent Systems at The University of Sydney and director of research and innovation at ACFR, that future’s not far off.

Professor Sukkarieh and an ACFR research team of 10 are developing robotic systems with the capacity to map and monitor farmland, gather information continuously, provide accurate yield counts, advise on optimal harvesting times and perform routine tasks such as pruning, weeding, mowing and spraying.

Don’t expect automated harvesting for about a decade; but Professor Sukkarieh thinks viable ‘intelligent’ farm systems could be just a few years away.

“Robotics is offering a pathway to reduction in labour costs, more efficient operations and better land management approaches, with significant potential to improve productivity,” he says.

Add sophisticated sensors and machine learning and there’s further potential to transform on-farm operations.

“Linking robotics with farm-wide sensing systems, large-scale data analytics and new machine learning algorithms is providing unprecedented levels of information about the real-time status of the farm and its overall health,” says Prof. Sukkarieh.

“Optimisation algorithms and automated statistical decision techniques are starting to make their way through into these systems, providing whole-farm management support tools that feed their way directly to the robot for real-time operation.”

One example is ACFR’s tree crop information processing project, which uses machine learning, large data analytics and robotics collectively to build large-scale farm information management systems, in real time and with high precision.

Professor Sakariah’s team is also employing machine learning and data analytic techniques to detect and identify invasive plant and animal species, building robotic aircraft and combining the data gathered with output from intelligent sensors to build high-resolution maps of farmland.  These maps can be used to assess various tree-crop characteristics, distinguish native vegetation and determine where to spray for weeds – potentially saving the agricultural sector around $4 billion a year in weed control.

Then there’s the ‘Ladybird’ project, in which the focus is on building a new robot for crop monitoring and weeding with the objective of sustainable operations: solar-powered and. requiring no herbicide use.

All these projects are funded by HAL, Ausveg, the Departments of Primary Industry in NSW, Queensland and Victoria, and the Australian Government. 

Further along the supply chain, field robotics and intelligent systems will likely be instrumental in helping farmers meet the increasingly rigorous standards of supermarkets and consumers.

“If you can send a robot out every hour, process the data and look at the changes, you … can adapt a lot more easily to new demands driven by regulations, new health specifications or consumer requirements,” notes Prof. Sukkarieh.

Given Australia’s high labour costs, regulated environment and reputation for ‘clean food’, export profits are most likely to flow from catering to the growing global demand for safe, premium produce. Here, too, robots and intelligent systems are set to play a key role.

“Asian growth …has shown that consumers are concerned about food quality and nutrition,” he says. “That’s where robotics could succeed, because [by using them], an Australian farm can become a lot more adaptable regarding export market demands. If the market changes, robotics can deliver a much better understanding of the farm and enable rapid decision-making.”

With regard to ROI, “It’s still early days,” says Prof. Sukkarieh, “although indications are suggesting [these innovations could be viable in] two to three years. The question is …how to quantify the benefits.

“It’s more than just saying, ‘The robot will remove the labour cost associated with weeding’ because there are also benefits associated with 24/7 operation, better energy efficiency, better overall understanding of farm health.”

For more information, visit www.sydney.edu.au/acfr/agriculture