Robotics: “Winemakers are unlikely to be out of a job anytime soon”

ow would you describe use of robotics in the wine industry?

In the wine industry, there is very little robotics or robot-related technology being used in commercial application, the only exception being bottling plants and associated warehouses which are highly automated. An increasing number of research projects are testing various mobile systems in vineyards.

Would you say that Australia is not as advanced, as advanced or more advanced than other countries when it comes to the use of robots?

In general, Australia is among the most advanced countries, although in the wine industry there aren’t sufficient worldwide use cases to make a comparison.

Can you quote any countries that you think are particularly advanced in the use of robotics?

USA, Japan, Germany and China

Do you think the industry is on the cusp of widespread usage or will it take a few more years?

It will take a few more years for widespread adoption of robotics, due to the inability to demonstrate benefit to farmers, training of operators and service providers for maintenance, need for improved robustness to environmental conditions (dust, rain, spiders, snakes, …), and to some extent the capital cost. As they become more widespread, issues such as liability, autonomy and trust will need to be dealt with.

Which areas of the industry would you say are the most suited to the use of robots?

In vineyards: mowing/mulching/slashing, crop monitoring, variable rate spraying

In wineries: oak barrel refilling, logistics

It seems that most efforts so far have been focused on the vineyard. Would you agree?

Yes, as mechanisation of the winery processes according to common manufacturing process improvement is readily available. This doesn’t necessitate the use of robots, or what are traditionally thought of as robotic systems. The large amount of vineyard passes on traditional machinery is a logical target for automation, hence effort has focussed on this.

How would you describe usage in the winery?

It is focussed on logistics, moving cartons and bottles around and filling them.

Do you see any barriers to rolling out more robots inside the winery?

No, only where they are needed.

Do you think use of robotics, either in the vineyard or the winery, can improve the quality of the wine or is it more about less/better use of human resources?

Improved sensing and control systems may improve the quality of the wine, but it is quite subjective anyway, so winemakers are unlikely to be out of a job anytime soon. Robots are more likely to help in reducing monotonous jobs.

Are there potential cross-overs from other industries that could be capitalised on by the wine industry?

There is substantial research going into the horticulture industry which is readily applicable in viticulture. As the uptake of autonomous cars increases, similarly equipped farm vehicles will rapidly become the norm.

At UNSW, what is your research focused on primarily?

At UNSW we develop algorithms and software systems for sensing in viticulture and horticulture from a robotics perspective. For industry we have provided solutions for variable rate spraying, low-cost mapping of canopies, an app for measuring vine water stress, yield estimation (shoot and flowering stages), microscope image based stomate processing and image stitching. We collaborate with other research institutions to provide image based solutions for berry counting, bunch architecture detection, disease detection, hyperspectral spray deposition analysis, non-destructive maturity estimation and automated sample naming. Furthermore we have developed a proof of concept autonomous concrete cutting robot and support a student-led project to develop a team of cooperative UAVs and UGVs.

What have been your achievements so far, and what will you be focusing on in the future?

We have demonstrated the possibility of estimating yield visually at the shoot and flower stages; developed and beta-tested an app for measuring vine water stress; developed robust algorithms for stomate detection; and we will be continuing development on all aspects of vine and tree sensing and modelling to provide actionable insight for farmers in a cost-effective manner.