Robot dawn in central Queensland

GroundCover Live and online, stay up to date with daily grains industry news online, click here to read more
Photo of Andrew and Jocelyn Bate at proof-of-concept trials

Andrew and Jocelyn Bate look to the future of grain growing with an early vision of a prototype agricultural robot undergoing proof-of-concept trials on their family's property near Emerald, Queensland.

PHOTO: Clarisa Collis

For Central Queensland growers Andrew and Jocelyn Bate, the era of farm robots has well and truly arrived.

Announcing its arrival are beams of infrared light moving slowly across the couple’s cropping area, revealing where an agricultural robot is seeking out and killing weeds.

The Bate family has been hosting proof-of-concept trials for a futuristic farmhand called AgBot for about two years on their 4000-hectare property at Gindie, Queensland, 20 kilometres south-west of Emerald.

Currently working to kill weeds on their fallow country, the autonomous machine, or AgBot, is the product of a $1.8-million research project that the couple initiated and partly funded.

The research project includes Queensland University of Technology, the Australian Centre for Field Robotics and the Bates’s own company, SwarmFarm.

Speaking at a GRDC Grains Research Update at Warra, Queensland, Andrew said the prototype is the forerunner for a much larger vision that he has to revolutionise grain growing.

AgBot is the prototype for a planned coordinated fleet of lightweight, low-cost robots that work together, using two-way radio communications, to perform a range of mechanised tasks such as spraying, spreading, seeding, harvesting and data collection.

Grounding this science-fiction concept are three imperatives: reduced expenditure on machinery; reduced input costs for herbicides, fungicides, fertiliser and seed; and reduced soil compaction.

Photo of AgBot

The AgBot's new linkage bar design.

To meet these specifications, the prototype has been developed using basic hardware and inexpensive software.

For example, the AgBot uses a $200 camera lens to identify obstacles and navigate without using expensive real-time kinetic GPS systems. Weeds are selectively sprayed using WEEDit® optical sensor technology.

Since the research began in 2013, Andrew said the AgBot has evolved from an electric golf buggy into a more utilitarian machine – a self-powered, automated, three-point linkage bar.

The linkage bar design also means the robot can be coupled to other farm equipment and adapted to a range of mechanised tasks.

For example, attaching a weed wiper to the self-powered AgBot allowed the couple to crop-top weeds across part of their 1500ha chickpea crop in 2014.

Weighing about 200 kilograms, compared with the 21-tonne burden of a conventional sprayer, also means the AgBots can venture into wet paddocks immediately after rain to spray weeds, allowing for more efficient control and. potentially, less herbicide use.

The slow speed of the prototype, which travels at up to 15 kilometres per hour, is another design feature expected to improve the precision of weed spraying.

In contrast with high-speed sprayers, the AgBot’s slow speed makes it capable of managing individual plants, as opposed to requiring large plant populations to target.

“Shielded sprayers and weed wipers are not practical on a conventional sprayer that travels at 28km/h,” Andrew said.

He said the projected reduction in chemical inputs means reduced environmental impacts and reduced selection pressure for herbicide resistance. The research has also been exploring the possible use of chemical alternatives that can take advantage of the slower operating speeds.

“Alternative approaches to weed control that target individual plants, such as microwave and steam technologies, could be grafted onto the prototype’s existing capabilities,” he said.

Andrew said another benefit likely to result from a gang of robots was the ability to minimise downtime caused by machinery breakdowns.

When equipment fails, relying on a single machine generally sees grains operations grind to a halt, whereas using several small robotic units could reduce or avoid downtime.

In the event that an AgBot might become bogged in the paddock, other robots in its gang can pick up the slack – sharing its workload and even navigating around the inactive machine, he said.

Andrew said the emphasis on technology that allows the robots to function as a cooperative swarm has already led researchers to develop a “docking and filling” system that enables the AgBots to automatically refuel.

The research partnership has opted to power the prototype with diesel because it is more cost-effective and less bulky than electric and solar alternatives.

With trials to test the prototype’s performance in different farming systems and growing regions starting in early 2015, Andrew hopes the AgBot will be available commercially for weed control as early as 2016.

More information:

Andrew Bate,


Biosensors to detect trouble early


Push to increase update of CTF

Region North