Case Study 1 – Beefwood Farms, the trials and tribulations of an automated driverless system
Located between Goondiwindi (Qld) and Moree (NSW), Beefwood Farms is an aggregation of six neighbouring properties serviced from a single workshop area. Gerrit and Pam Kurstjens purchased the property in 2006 and now run a 12,000ha operation. Having seen great advantages in optical spot spraying over the past ten years, Gerrit was looking to achieve greater efficiencies in fallows. Gerrit said ‘sourcing labour is extremely difficult and expensive; the spot sprayers can be filled and then need to be moving around the clock for weeks on end to cover our entire property. For this reason, we could see a lot of efficiencies in an autonomous system’. Gerrit and the team at Beefwood are one of very few commercial systems in the world to adopt an autonomous tractor system in their farming business.
‘Precision Makers’, a Dutch company had developed an autonomous, driverless system for lawnmowers, and were able to modify it and install the necessary software to fit to the Fendt 936 Vario tractor that Gerrit currently owned. Precision Makers also went on to fit the system to a John Deere 8345 tractor purchased for the farm. Gerrit said ‘we were seeing massive advantages with the autonomous technology; we could set spot spray system up and essentially go home and monitor it. Autonomous machines became an economic solution for work that is often considered uneconomic. The team at Beefwood were also doing their CTF (Controlled Traffic Farming) tramline track renovating with an autonomous tractor’.
In 2019, Precision Makers were bought out, and no longer supported the autonomous system in broadacre equipment, ‘we are desperately looking for another company, ideally closer to home that can service and support our machines’ said Gerrit. Dr Rohan Rainbow, industry colleague said ‘the tyranny of distance from Europe to maintain the equipment at Beefwood, the lack of mobile phone connectivity for remote support, lack of tractor firmware and ISOBus compatibility, plus the legal risk of operation without any regulation or an industry code of practice were all issues’ providing roadblocks. Gerrit, along with many Australian farmers has received the draft of the Autonomy Code of Practice and will have the opportunity to provide feedback’ he said, ‘we hope this is one step towards Beefwood Farms once again using the autonomous technology of which they have seen great value’.
Case Study 2 – A New Concept for Flower Thinning in Apple Orchards
SwarmFarm is one of the few businesses in the world that has delivered autonomous robots to farmers, with over 200,000 acres commercially sprayed, mowed or weeded with their robots. The SwarmBots developed by SwarmFarm are capable of being adapted to a variety of different farm functions, and can work independently, or ‘swarm’ in unison with other SwarmBots. This adaptability is what makes them ideal for an autonomous carrier of different technologies. This is why they have developed a partnership with ‘Green Atlas’ – the hardware and software provider making exciting breakthroughs in apple orchards.
The SwarmFarm team has developed a new concept for apple orchards which is referred to as ‘VTA’ – Variable Timing of Application. The SwarmFarm project (in collaboration with partners - University of New South Wales (UNSW), Adama, Apple and Pear Australia Ltd (APAL) and Hort Innovation) is being carried out in apple orchards of Batlow, NSW and is about identifying the flowering stage of individual trees with the concept of applying thinning products at the optimal time for each individual tree within the orchard.
Apple trees are variable, each year they have different volumes of flower load, leading to variable fruit yield. Growers in the market are looking for a more automated and targeted approach to flower thinning regimes, to help overcome the challenges and costs of sourcing labour for hand thinning. The other challenge for growers is getting the timing of the application of the thinning chemical correct because the length of the flowering window is also highly variable within an orchard.
With these pressure points in mind the team is developing a camera system that drives down each row photographing each tree to generate a ‘flowering stage’ map. This has been possible by training an algorithm comparing the data collected from the cameras with flowering assessments from specific areas of the field. So far, the mapping is showing ‘amazing accuracy’ according to Dr Mark Whitty. These maps are processed and analysed to produce an application map that targets only specific trees within the orchard that need thinning on the day of application. So rather than treating the entire orchard at the same time, varying timing can target individual trees and apply primary chemical thinners at the correct growth stage.
2020 is the third season data has been collected in the orchard. There is great confidence already in detecting flowering from green tip to petal fall, so those working on the project are now looking to be able to detect variability at the fruitlet stage. Dr Mark Whitty says ‘we can provide data very close to real time now’; as opposed to when they first started - an offline process taking many hours. The variable flowering maps are also showing benefits to growers when seeking agronomy advice, with agronomists able to make comment on spraying decisions without having to visit the orchard.
For more information visit: http://www.robotics.unsw.edu.au/srv/project/apple-sensing.html
Case Study 3 – Milking edge: a training and extension program for automatic milking in Australia
The Australian Dairy Industry is committed to supporting investment and successful operation of Automatic (Robotic) Milking Systems (AMS). ‘Milking Edge’ is a three-year project as a collaboration between the New South Wales Department of Primary Industries, Dairy Australia and DeLaval to support industry to invest and operate AMS successfully.
The first robotic dairy technology came to Australia in 2001, having been adopted in the Netherlands. Since then, over 50 farms in major dairy regions have invested in AMS technology. In 2004, the collaborative project ‘Future Dairy’ was initiated, driven and funded by industry and in 2006 established the first automatic milking research dairy at Camden, NSW. ‘Future Dairy’ was responsible for Research and Development of automatic milking systems, looking into how Australian dairy farmers can successfully implement AMS systems in pasture based situations, with clear benefits for cows and people.
Built on a decade of successful research and development led by FutureDairy, as well as the experience of almost 50 commercial farms around Australia who have decided to invest in the technology, the Milking Edge project will enable better decision-making around the consideration, purchase and implementation of AMS.
One of the outcomes of the project is the development of eight online training modules, three of which are already available, and four that are close to release. The modules are free, and cover topics such as infrastructure, labour, economics and reproduction management.
The project is spearheaded by Nicolas Lyons, Jess Maloney (Department of Primary Industries) and Juan Gargiulo (University of Sydney) who will work closely with key stakeholders in the dairy industry in both Australia and overseas. Nicolas Lyons says ‘we are confident that the extension from this project is helping greatly with the successful adoption of AMS in Australia. We have a closer working relationship with 65% of farms with the technology. Into the future, we hope to be able to work on utilising the 120 data points collected at milking time to further enhance and automate the milking process’.