ROBOTICS
AND INTELLIGENT SYSTEMS: THE KEY TO FUTURE GROWTH IN FARMING
Recent
discussions about Australia's potential to become the 'food bowl' of south Asia
have highlighted both the opportunities and the challenges involved. Will we
have the capacity to meet demand and make the most of the looming agricultural
boom? The Australian Centre for Field Robotics is playing a key role in
developing robotics and intelligent systems to support this growth and help us
rethink the future of farming.
A trial involving ground and aerial robots on an almond farm
in Mildura
Imagine
a tractor that can travel, driverless, up and down the rows of trees in an
orchard, 'observing' its surroundings and developing a detailed 3D map that
shows the position and condition of every tree, every fruit and every inch of
ground. Imagine it can identify any disease or deficiencies in each tree, the
ripeness of each fruit and the condition of the soil. Then imagine it can
respond by spraying, watering or harvesting as required - all without the need
for a human operator on board. This is the farm of the future, and it's already
being developed at the Australian Centre for Field Robotics (ACFR).
The ACFR develops 'intelligent' robotic devices for use in outdoor environments, which can autonomously sense, analyse and respond to their own surroundings. Professor of Robotics and Intelligent Systems Salah Sukkarieh and his team are currently working on a project whose aim is to automate many of the previously labour-intensive operations of farming, vastly improving efficiency, yield and worker safety.
Professor Sukkarieh explains: "There's a big drive at the moment to conceptualise the future of Australian agriculture in terms of a 'food bowl' supplying the vast Asian market. The Asian population is booming, but the region's relative lack of arable land, water, infrastructure and technology restrict its ability to meet the associated increase in demand for fresh produce. Australia, with its reputation for high-quality produce, therefore faces a potential opportunity to ride another major economic boom - this time in agricultural rather than mineral resources.
"However, currently we don't have the capacity to do this to the degree that we could. One reason for this is a lack of labour - there just aren't the numbers of people we'd need who are both willing and appropriately skilled to work on farms. Another is the cost of labour - even if we did have the numbers, the cost would be prohibitively high. The other major reason is that, like the wider Australian population, the average age of our farmers is rising, so they're becoming less able to perform the hard physical labour traditionally involved in running a farm.
"This is where automation can help. We can use it to increase efficiency and yield, by having many of the manual tasks of farming performed by specially designed agricultural robotic devices."
The project involves three stages, the first of which - using 'autonomous perception' to enable robotic devices to 'read' and 'understand' their surroundings - is currently underway and will be made commercially available to farmers within the next couple of years. The ACFR team, with the support of Horticulture Australia, has developed robotic systems, sensors and intelligent devices such as those shown here on an almond farm in Mildura, which can move through an orchard gathering data and developing a comprehensive 'in-ground and out-of-ground model' of the entire orchard.
The ACFR develops 'intelligent' robotic devices for use in outdoor environments, which can autonomously sense, analyse and respond to their own surroundings. Professor of Robotics and Intelligent Systems Salah Sukkarieh and his team are currently working on a project whose aim is to automate many of the previously labour-intensive operations of farming, vastly improving efficiency, yield and worker safety.
Professor Sukkarieh explains: "There's a big drive at the moment to conceptualise the future of Australian agriculture in terms of a 'food bowl' supplying the vast Asian market. The Asian population is booming, but the region's relative lack of arable land, water, infrastructure and technology restrict its ability to meet the associated increase in demand for fresh produce. Australia, with its reputation for high-quality produce, therefore faces a potential opportunity to ride another major economic boom - this time in agricultural rather than mineral resources.
"However, currently we don't have the capacity to do this to the degree that we could. One reason for this is a lack of labour - there just aren't the numbers of people we'd need who are both willing and appropriately skilled to work on farms. Another is the cost of labour - even if we did have the numbers, the cost would be prohibitively high. The other major reason is that, like the wider Australian population, the average age of our farmers is rising, so they're becoming less able to perform the hard physical labour traditionally involved in running a farm.
"This is where automation can help. We can use it to increase efficiency and yield, by having many of the manual tasks of farming performed by specially designed agricultural robotic devices."
The project involves three stages, the first of which - using 'autonomous perception' to enable robotic devices to 'read' and 'understand' their surroundings - is currently underway and will be made commercially available to farmers within the next couple of years. The ACFR team, with the support of Horticulture Australia, has developed robotic systems, sensors and intelligent devices such as those shown here on an almond farm in Mildura, which can move through an orchard gathering data and developing a comprehensive 'in-ground and out-of-ground model' of the entire orchard.
The view from the aerial robot
"Traditionally
it has been necessary for someone to actually walk through the orchard, taking
and analysing soil and other samples and making decisions on the health and
yield quality of the plants," Professor Sukkarieh says. "The devices
we've developed can collect, analyse and present this information autonomously,
so a major part of the farmer's job can be done automatically."
The second stage, which the team will work on next, involves applying this technology to standard farm tractors, so that as well as being able to perceive their environment and identify any operations required, they will also be able to perform many of these operations themselves, such as applying fertilisers and pesticides, watering, sweeping and mowing. The third and most complex stage will be to enable the devices to carry out the harvest.
"Perception is the fundamental quality of all autonomous systems, and the key to overall capability," Professor Sukkarieh explains. "For this reason we've tackled that stage first. Once that part of the technology is established, we can work towards more interactive tasks such as autonomous harvesting. The reason this is more complex is that it needs to take into account the particular type of crop.
The second stage, which the team will work on next, involves applying this technology to standard farm tractors, so that as well as being able to perceive their environment and identify any operations required, they will also be able to perform many of these operations themselves, such as applying fertilisers and pesticides, watering, sweeping and mowing. The third and most complex stage will be to enable the devices to carry out the harvest.
"Perception is the fundamental quality of all autonomous systems, and the key to overall capability," Professor Sukkarieh explains. "For this reason we've tackled that stage first. Once that part of the technology is established, we can work towards more interactive tasks such as autonomous harvesting. The reason this is more complex is that it needs to take into account the particular type of crop.
For
example, to harvest almonds you can simply shake the tree and let the almonds fall
to the ground, then sweep them up. But for other tree crops such as apples,
bananas or mangoes, you need a manipulator arm that's appropriate for that
crop, that can individually pick only the ripe fruits, that won't bruise them
and that can pack them appropriately.
"The devices we've developed already can identify each individual fruit on the tree and its degree of ripeness, which is about 80 percent of the job done. But being able to harvest them is our ultimate goal."
Not surprisingly, Professor Sukkarieh reports that the project already has strong support from industry: "There is significant interest in the agricultural community about the applications of robotics, which is really encouraging. As project leader I've been invited to address several grower conferences for various fruit and nut industry bodies about the implications of this technology to the future of farming."
"The devices we've developed already can identify each individual fruit on the tree and its degree of ripeness, which is about 80 percent of the job done. But being able to harvest them is our ultimate goal."
Not surprisingly, Professor Sukkarieh reports that the project already has strong support from industry: "There is significant interest in the agricultural community about the applications of robotics, which is really encouraging. As project leader I've been invited to address several grower conferences for various fruit and nut industry bodies about the implications of this technology to the future of farming."
As
well as developing the technology itself, the ACFR team is working with farmers
to determine how making small changes to traditional agricultural practices can
allow them to make the most of this new technology.
"For example," Professor Sukkarieh says, "growing apples on trellises makes it easier to apply automation to the full extent possible. Relatively small changes like this can generate significant increases in efficiency, so we're actually looking at designing the next generation of farms to work with this kind of autonomous technology."
The ACFR is already recognised internationally for its world-leading research, development, commercialisation and theoretical contributions to field robotics. One of the largest robotics research institutes in the world, its previous projects have included a pilotless aircraft that can travel over remote bushland identifying invasive weeds and delivering targeted applications of herbicide; other intelligent aircraft for use in border patrol, military operations and management of livestock, feral animals and native fauna; and intelligent underwater, land and space vehicles with applications in marine science, mining and planetary exploration.
This latest project will further enhance its reputation and enable Australia to become what Prime Minister Julia Gillard has described as a prospective 'food superpower' in the region.
"For example," Professor Sukkarieh says, "growing apples on trellises makes it easier to apply automation to the full extent possible. Relatively small changes like this can generate significant increases in efficiency, so we're actually looking at designing the next generation of farms to work with this kind of autonomous technology."
The ACFR is already recognised internationally for its world-leading research, development, commercialisation and theoretical contributions to field robotics. One of the largest robotics research institutes in the world, its previous projects have included a pilotless aircraft that can travel over remote bushland identifying invasive weeds and delivering targeted applications of herbicide; other intelligent aircraft for use in border patrol, military operations and management of livestock, feral animals and native fauna; and intelligent underwater, land and space vehicles with applications in marine science, mining and planetary exploration.
This latest project will further enhance its reputation and enable Australia to become what Prime Minister Julia Gillard has described as a prospective 'food superpower' in the region.
Professor Salah Sukkarieh
Research Director
Australian Centre for Field Robotics
The University of Sydney
Research Director
Australian Centre for Field Robotics
The University of Sydney
