As the Australian Government’s skills summit approaches, the UK agriculture department has published an expert report on the potential to automate and transform labour productivity in the horticulture industry.

In 2019-2020, Australia’s horticulture industry - comprising fruit, vegetables, nuts, flowers, turf and nursery products - exceeded $15 billion in production value and exports were over $3.1 billion. Yet with an estimated shortfall of 25,000 or more seasonal workers, produce in some areas is withering before it can be picked and crops are dying before they can be harvested.

The UK horticulture industry is also facing a labour crisis, with COVID compounded with labour from Eastern Europe drying up post-Brexit.

The unique challenges of automating horticulture

First, each horticultural crop grown has its own unique husbandry, harvesting and handling needs. Growing strawberries and other ground crops is a very different exercise to growing tree-based fruit. Horticultural crops also have distinct growing seasons and seasonal workers are employed at key times in different cycles. As a result, horticulture is a collection of unique sub-sectors with broad and diverse needs. Where other industry sectors have successfully automated, they have done so by removing variation and standardising components. However, as the UK report points out, variation is inherent in horticulture.

Second, infrastructure, horticultural practices - even the plant varieties themselves - have been optimised for human workers. The UK report stresses that there would need to be broader industry changes in what we grow and how we grow it to optimise the industry for automation and robotics.

Third, horticultural farms tend to be small family-run operations: in 2019-2020, there were over 2,850 horticulture farms in Victoria alone. Automation is capital intensive, and will stain the financing capacity of many current producers.

Finally, while seasonal worker jobs in horticulture are considered low-skilled, they require considerable human stamina, cognition and dexterity. Robots are still struggling to match the dexterity of human hands in harvesting fruit:

“In orchards with trellised trees, human fruit pickers can cruise through rows of trees in pairs on slowly rolling platforms. One person crouches to reach low-hanging fruit, the other reaches for the higher branches. Professionals working this way take about two seconds to pick one apple. [A prototype robot picker] is essentially a giant arm mounted on a rolling platform, takes about five seconds to make its moves. At the click of a key, the robotic arm reaches up for the fruit…with its three-fingered palm. Its fingers are covered in cushiony silicone “skin”, which conceals individual motors wired to tendons that drive its fingers. Thirty sensors under each fingertip track the pressure, speed, angle and other aspects of its grasp to help the robot complete its task. Another keystroke and the fingers tighten, then twist, and the apple – successfully picked – rests in the robot’s palm.”

The automation time horizons for horticulture

The report provides a realistic timeline for a staged automation transition in horticulture – which then shapes its recommendations.

As figure 1 shows, the report identified three ‘waves’ of automation:

Wave 1: next 3-5 years

There are three ‘clusters’ of technologies available to the horticulture industry now or in the near term:

• Optimised production systems: such as improved infrastructure like more energy efficient greenhouses, infield canopy architectures over vines to protect against bad weather, and farm management systems.
• Packhouse automation: Packhouse automation performs low variation tasks in a controlled environment, such AI-based tools that learn to sort fruit by size, colour and surface appearances (e.g. the degree of blemishing).
• Field rigs and mechanical systems: used in field operations to partly automate some husbandry and harvesting tasks.

The labour savings from these wave 1 technologies are limited for two reasons: most labour is in the field and not in the packhouse, and the automated assistance in the field is mainly ergonomic and helps improve worker efficiency (and reduce physical demands), but does not replace them.

Wave 2: closer to 2030

• Augmented work: these systems, including those enabled by Artificial Intelligence (AI) or collaborative robotics, improve a worker’s productivity through technology, often depending on a worker’s cognitive abilities and dexterity for the more complex aspects of a task. Examples include exoskeletons to support carrying heavier weights of picked fruit or scaling trees, or wearable technologies, such as a vacuum glove for berry picking. While enhancing efficiency, they do not replace seasonal workers and will require some retraining of those workers to optimise the technology.
• Autonomous crop protection, monitoring and forecasting: this suite of autonomous systems (such as robotic sprayers, light treatments or fruit counting) helps underpin the transition towards full crop harvesting robotics and are considered critical technical steppingstones. These technologies also could be available before 2030, but offer low labour savings at a sector level.

Wave 3: beyond 2030

• Autonomous selective harvesting: mobile robotic systems that autonomously navigate growing environments, identifying and harvesting crops (e.g. picking apples). Autonomous selective harvesting offers high labour savings at a sector level but if left to market forces will unlikely be commercially available until well after 2030.

Many of the waves 2 and 3 technologies are currently in development pipelines, and it is not yet clear whether they will survive the ‘valley of death’ faced by most new technology in early stages due to the large capital required for development activities.

Recommendations

The UK report’s recommendations were built around 3 key themes of 'mind the gap', 'collaboration' and 'technology alone is not a solution'.

Mind the gap

The key implication of the UK report is that there is a need for a secure source of labour in the period before mass-adoption of automation technology is feasible. As the following graph shows, the real savings in field staff are at the ‘back-end’ of the transition process to a more automated industry:

But equally as important as the seasonal workers themselves is a stable, multi-year Government commitment to a seasonal worker program to justify the future investment in technology: “growers were also concerned whether the policy environment, specifically the future of the UK’s seasonal worker program], would have sufficient scale and duration for them to confidently bridge the transition towards greater automation.” The length of any future schemes should ideally match the period preceding the feasible mass-adoption of automation technology.

Collaboration

The UK report also considered that market forces alone would not be enough to get automation out of the starting blocks because of the fragmentation of the horticultural industry. The report recommended the Government convening a consortium that brings together UK government departments, horticulture industry and technology companies with an agenda to:

• fast-track the adoption of available technologies across horticulture, and the transfer of proven technologies across sub-sectors
• increase collaboration across the supply chain
• increase knowledge transfer and the sharing of risk at all stages of adoption.

Both to create an energising catalyst for change and to bring forward technology with the most labour-saving capacity, the report recommended the Government launching a robotic crop harvester mission to fast-track innovative research and development.

In what is probably an overly optimistic reading of technology projects backed by Governments, the report also considered that a more proactive technology development approach by the UK Government could promote the emergence of standardised platforms. Growers cannot invest in multiple systems for multiple tasks or crops due to the high cost of technology. A common platform that sub-systems can be added to would be more economically viable for growers.

Technology alone is not a solution

The UK report makes the obvious but good point that technology cannot be successfully adopted and exploited without the appropriate supporting mechanisms: chief amongst these are improved connectivity of communications networks in regional and rural areas and new digital workforce skills.

As a recent digital agriculture report in Australia also identified, horticulture, and agriculture in general, face a ‘wicked’ problem in finding and retaining the STEM skills needed to support transformation:

• agtech sector needs to compete against higher paying, urban-based economic sectors for STEM workers;
• those STEM skills needed to support horticulture need to be locally available in regional and rural areas:

“Sometimes technology fails, and it is not commercially viable for a system to be down for a long period of time when a grower's window for harvest is short. Where problems occur, a skilled individual needs to be on site within a matter of hours. This requires maintenance services to be local, abundant and on-demand in a sector that is geographically dispersed.”

We are not talking about the high level of STEM skills consumed by the Atlassians of the world. As the Australian agritech report discussed, more focused, locally TAFE-based skills courses may provide an avenue to build up local tech ecosystems in regional and rural areas.

The report also recognised new economic opportunities for SMEs supporting the transformation of horticulture. Automation integrators are technology specialists that combine off the shelf technologies to build an overall bespoke system. They normally work across multiple sectors and, once they have created a system, often exploit the concept by selling the system to others.

There are challenges in attracting the attention of integrators to horticulture. The specialised nature and scale of horticultural subsectors means the potential market for building a new system is small. Additionally, the pool of UK integrators is in high demand across all sectors, so they might decline more technically challenging projects. There is no central information point signposting growers to integrators who will work on horticultural commissions, which means many growers have lost momentum in adopting automation.

Finally, there also may need to be a commitment by large wholesale purchasers, such as the supermarket chains, to transformation in horticulture. Automation has a payback period over many seasons, so very few growers will commit to long term investments when the future of their business is uncertain. Due to uncertainty in the duration, security and scale of production contracts, growers can lack the confidence to invest.

Takeouts for the Australian skills summit

The UK report is clear that robotics and automation technologies are not likely to be the single solution that resolves all labour pressures in the sector. However, the review, as the Australian agritech report before it, seeks to articulate a series of recommendations for the government to consider that could have the power to transform labour productivity over the medium to longer term.

Australia long has been a leader in agricultural technology, and there are some good examples in horticultural technology:

• ‘Clive’ a fruit picking robot has been deployed in the Goulbourn Valley;
• The Australian Banana Growers’ Council has developed a mobile app to detect and report bunchy top disease;
• Thingc Robotics is developing low cost robots to undertake horticultural work, such as weed spraying.

The key messages from the UK report are ‘start now, view problems and solutions through an SME lens, be more co-ordinated industry-wide, and build locally’. Food for thought for the upcoming skills summit.

Read more: Automation in horticulture review