The Federal Government has put $2.9 million into a multi-partner development that aims to revolutionise the labour-intensive business of managing weeds on farmland.
Subscribe now for unlimited access.
$0/
(min cost $0)
or signup to continue reading
The "Kelpie" autonomous robot is being developed to selectively control weeds, including in pasture or vineyards - addressing the "green on green" problem. The project aims to provide farmers with an attractive alternative to labour-intensive spot spraying or boom spraying.
The robot will use custom-built algorithms to identify and control individual weeds, either through spraying or digging.
The University of New England (UNE) will play a significant partnership role in the Cooperative Research Centre Project (CRC-P), providing its networked SMART Farms for development and testing and making available its research into plant identification technologies.
UNE will also provide the project with a post-doctoral researcher and an engineer, to address the challenge of recognising weeds as the Kelpie approaches them.
READ ALSO:
The project is led by software and robotics company Agent Oriented Software (AOS), whose managing director, Dr Andrew Lucas, forecasts that autonomous weed control will offer several practical advantages to farmers.
"The Kelpie platform doesn't require a human to operate it, so it reduces labour costs; its selective action avoids the collateral damage to pasture that occurs with boom spraying, and it minimises chemical use," Dr Lucas said.
"As a bonus, the Kelpie will photograph and record the variety and positions of the weeds, providing the farmer with a 'weed map' allowing annual assessment of control efforts and providing verification of weed control measures taken.
"The Kelpie will also assess pasture quality, to provide a feed quality map that interprets the effect of weeds on productivity.
"The Kelpie will be based upon a proven agricultural robotics platform, capable of withstanding Australian agricultural conditions, and incorporate the latest sensors. The aim is to have the first pre-production Kelpies in the field in 2022."
NSW Department of Primary Industry is a key partner in the project, and the library of thousands of photographs of weeds and grasses taken during Kelpie's development will be made available in a Reference Weeds Library as a research resource.
Leader of the successful project bid team, UNE's Director, Strategic Research Initiatives, Associate Professor David Miron, said the university is delighted to be contributing to such an important project.
"Weed management in crops is well advanced, partly because crops are grown in regular patterns," Professor Miron said.
"In grazing situations, weeds occur randomly. In these environments, we still have to deal with weeds manually, which is often not time- or cost-effective.
"The Kelpie project draws on the extraordinary developments we have seen in robotics, remote sensing and AI over the past few years, and puts them together in a package that we hope will completely change the economics of managing weeds outside cropping environments."
Federal Member for New England Barnaby Joyce said the government's $2.9 million investment in the project would further support the region's pioneering traditions.
"Armidale and the New England have a long-standing reputation as a leader in innovation, especially in agriculture," he said.
"This new technology to manage weeds will complement that and ultimately deliver a better return through the farm gate."
The "Kelpie" Cooperative Research Centres Projects (CRC-P) grant was one of 10 to be granted Federal funding in the latest CRC-P round.
The project, fully titled "Novel autonomous robotic weed control to maximise agricultural productivity", is valued at more than $8 million when the contributions of its partners are taken into account.
The project's partners include LF Autonomous Agriculture, AOS's commercialisation arm; Queensland University of Technology, which has specific expertise in Artificial Intelligence; Monaro Farming Systems, whose 71 members collectively manage 124,000 hectares of livestock enterprises; Treasury Wine Estates, which will support the project's development for vineyard use; and the NSW Department of Primary Industries.
Five Kelpie systems will be trialled on Monaro farms and Treasury Wine Estates, with the objective of autonomously navigating pastures and vineyards to economically control weeds where it this is currently not possible.