Robots herd cattle and pick broccoli


Humans have been farming for thousands of years since the first crops were planted and animals began being domesticated during the Mesolithic era. But technology could soon be squeezing us out of our traditional roles as farmers, as robots take to the fields. Source: Daily Mail UK

A robot that is capable of herding cattle and pulling trailers through the mud has gone on trial at a farm in Australia while another capable of pick broccoli six times faster than humans is being tested in Britain.

SwagBot, the Australian robotic cowboy, was built to be a low cost, powerful robot to work across the expansive, rugged areas of Australian farm terrain.

The machine has been showing off its skills in a trial that began last month at a farm near Newcastle, Australia.

The trial revealed SwagBot can herd cattle, and navigate its way around ditches, logs, swamps, and other features of a typical farm landscape, which it showed off in a YouTube video.

The battery-operated machine with all-wheel drive can reach speeds of 15 to 20 kilometres per hour on smooth terrain.

Over the next few years, the University of Sydney lab intends to build more autonomy into the robot so it can perform certain tasks such as animal monitoring or weeding, said Professor Salah Sukkarieh, who is leading the trial.

During the field test, the robot performed as was required and frightened cows.

“They were obviously scared and ran away. It’s what we expected,”Professor Sukkarieh said. “We use it to our advantage in herding animals, or we figure out how to monitor from a distance.

“Over the next few months, we’ll be looking at what algorithms we need to put together to allow the animal monitoring.”

At the same time in the UK, a robot could soon improve the efficiency of farmers picking crops.

The University of Lincoln has been developing the broccoli-picking robot, which can identify broccoli in fields with up to 95% accuracy.

The broccoli-harvesting robot uses a 3D camera taken from a Microsoft Kinect games console.

“Broccoli is one of the world’s largest vegetable crops and is almost entirely manually harvested, which is costly,” project lead Professor Tom Duckett, group coordinator of the Agri-Food Technology Research Group at the University of Lincoln said.

“This technology is seen as being an important move towards developing fully automatic robot harvesting systems, which could then be used for a variety of different crops.”

The researchers developed the imaging tools to allow the robots to identify the broccoli.

“This uses 3D cameras, basically Kinect cameras deployed on video games, to sense where the broccoli are located in the field,” Dr Simon Pearson, director of the Lincoln Institute for Agri-food Technology said.

“We are now working on plans to develop the next step which is the cutting mechanism. ‘However, we think we have cracked the imaging part, which is quite tricky,” he said.

“Ultimately with a robot it needs a 3D camera as you have to provide it with an x,y,z location of the head in the field.”

The camera needs to identify the exact location in all dimensions to tell the robot where to pick. Finding the depth, or the ‘z’ location, has always been the difficult part for the cameras, Dr Pearson said.

“We have focussed on broccoli as a starting point and think are aiming for a commercial system in 2 years.”

But Dr Pearson added the technology might be extendable to many fresh produce crops such as apples, pears, soft fruit, tomatoes, peppers and cucumbers.