Features
Is autonomous agriculture a premature promise?

Farming, open air or in greenhouses, is a stressful profession. Stressors include human error, which can cost money and jeopardize crops; unpredictability due to climate change and increasingly extreme weather; market uncertainty; debt; and a lack of time to accomplish tasks. Loss of skilled labour is another stressor. According to a 2022 report from the World Economic Forum, in Canada, the loss of skilled agricultural labour could increase to 114,000 by 2025.

To this list of concerns, Brian Lynch, PhD, Director, Horticultural Technology Solutions, Vineland Research and Innovation Centre, adds profitability. While other businesses in the greenhouse value chain, such as retailers, have leeway to accommodate increasing costs, Dr. Lynch notes that growers face smaller margins. Many of the issues growers face are connected to a lack of control that contribute to both mental anxiety and physical fatigue. Suicide is a risk for farmers due to the stressors they face (Alam, 2023).

Promoters of Autonomous Agriculture (AA) state that artificial intelligence (AI) applications will facilitate “inferences, predictions and decisions” and will allow agriculture to become “hyper-optimized,” according to Jacqueline Keena, managing director of Manitoba-based Innovation Farms that specialize in digital agriculture (Saba, 2024). AA supporters believe that it also offers growers precise assessments and speedy “interventions,” beyond human abilities (Saba, 2024), and the ability to use data from a variety of sources and give directions for future plantings (TTControl, 2024).

A cautious view of AA credits the application of technology to include AI in robots, as well as sensors, drones, software and digital twin computer systems for monitoring fertigation. AA is presented as a solution to farmers and growers for a variety of issues with an underlying promise of greater control and predictability with fewer costs. 

Nourish Labs, based in British Columbia, offers growers a fertigation management software system “designed to produce the most consistent, highest quality and yielding crops possible” (Nourish Labs). Their “nutrition formulations are “fresh, homogenized, PH-balanced plant nutrition formulations with milliliter precision to every plant at exactly the right time” (Nourish Labs). Nourish Labs offers growers control and precision. The pH balance remains stable with “uniformity” in plant feeding. The feeding itself is controllable by amount and time of application. The system increases ROI as it decreases effluent runoff, use of fertilizers, and use of water. It also decreases risk of human error, and with that is a matching decrease in worry, says Justin Valmont, Nourish’s chief product officer.

The company’s software is a “fifth-generation platform that is completely modular,” said Valmont, with a “base system” for growers to which modules can be added as needed. Nourish Labs’ technicians can visit onsite to assess a growers’ needs as well as to review the system’s electrical, plumbing, and irrigation needs. Their software system is used by greenhouse staff members who have “management responsibilities” such as Master Growers and Operator Managers. Access is available remotely and the system works 24/7. 

Can we trust robots to take our place in the greenhouse? Experts say trust is one of the major roadblocks to growers embracing high-tech innovations. Photo Courtesy Vineland Research and Innovation Centre

All that glitters is not gold

“Trust is a major issue when considering new technology,” notes Lynch. “Growers are bombarded with many pitches and claims by companies offering amazing new technologies and have to filter out which ones have real value.” 

Growers also have to consider ROI, performance, maintenance and support, and reliability. An issue with adoption of AA, Lynch observes, is that it is assumed that the current physical layout of greenhouses is what AA will be applied to. The future physical layout of greenhouses may be different. Additionally, greenhouse robots may have highly specialized functions such as watering, fertilizing, and pest removal. Alternatively, different functions may be found on one robot with different attachments. Plants may also change physically through breeding with smaller leaves and longer stems, for example, to ensure they can be successfully tended to by robots/automated attachments.

The curse of success

Lynch cautions that “how we grow plants evolves and issues plants face evolve too, such as climate change and types of diseases.” The risk in adopting AA, according to Lynch, may lie “in taking the risk today with technology that might not be ready or as high-performing as you want, versus waiting too long to adapt and having to catch up to the rest of the industry.” He foresees the potential for dependence on automation and AI to lead to future generations of growers with less intuitive knowledge and wisdom than their predecessors. 

“Growers have an intimate knowledge of their crops and years of experience and insights that make them successful. We’re now training AI systems to learn and replicate that expertise and seeing amazing results,” says Lynch. “But, while you can ask a grower to explain his or her reasoning and then pass that information to the next generation, it is often impossible to do so with AI systems.”

Safety, specialisation, certification

When it comes to robots, AA “is a bit optimistic, but not overly ambitious,” says Dr. Medhat Moussa, PhD, P.Eng., Systems and Computer Engineering, School of Engineering, University of Guelph. Innovations in technology, such as AI and sensors, that will eventually evolve into AA applications, begin as applications in cars and manufacturing. Some aspects of AA exist but for various reasons cannot be widely applied. Moussa notes that “autonomy” is one of the more ambitious elements of AA technology. Safety features and certification protocols used in industrial settings, such as cages for robots to work within or lasers to keep them on track, are being developed for robots to operate inside greenhouses. Expense is another issue. Hyper spectral cameras used to detect disease, for example, are still cost prohibitive for most growers as are robots that are specialised to provide their services to specific crops such as strawberries versus green peppers.  

The varying needs of growers also influences the implementation of AA in greenhouses, Moussa points out. The needs of a grower with 10 acres of greenhouses versus that of a grower with 200 acres are different, which affects AA adoption. The grower’s appetite for risk can also curtail the use of AA. Additionally, the power of AA such as is found in images taken of individual plants can provide “massive data,” Moussa notes, that exceeds the current abilities of most growers to work with that information. AA, Moussa believes, will have wider adoption in Canada in the next 10 years.

Nourish Labs, based in B.C., has developed a precision fertigation system that employs a fifth-generation platform that is completely modular. The system has shown to significantly reduce effluent runoff, use of fertilizers, and use of water. Photo Courtesy Nourish Labs

Cybersecurity, money, politics

Connectivity is a strength and a weakness. The strength of AI gathered information may be slowed by weaknesses in “connectivity infrastructure” when 4G and 5G connections may be needed (TTControl, 2024). Additionally, there is concern in the agricultural industry about the data new technology generates and the ‘bad actors’ that may find it worthwhile to steal or hold hostage. In the past three years Ontario Pork, Lactanet, a Canada-wide organization, Maple Leaf Foods, Agronomy Company of Canada, and JBS had ransomware attacks and data stolen (Martin & Rudolph 2024). Embracing AA means embracing data and computer system protections and restricting access to technology and information.

Current legislation is seen as inadequate to protect farmers and growers according to David Shipley, CEO of Beauceron Security (Connolly, 2024). Bill C-26 – An Act Respecting Cyber Security (2022) protects finance, telecommunications, energy, and transportation but did not include agriculture. It is hoped that amendments will enhance current protection for software and “smart” agricultural equipment.

ROI

Initial implementation of AA is likely the most investment heavy stage in the lifecycle of AA in a greenhouse. “The main challenge in automating planting and seeding lies in the initial cost and complexity of integrating these systems into existing operations. Additionally, customization may be necessary to accommodate different types of seeds or planting patterns, adding to the expense and complexity” (Bernier, 2024). While AA is sometimes presented as a panacea technology is not perfect — there can be failures. “Implementing these systems can be complex, as they require detailed programming to adapt to the varying needs of different plants. There’s also a risk of mechanical failures, which could lead to under or over-watering, potentially harming the crops” (Bernier, 2024).

Before investing in AA, it is advised that growers consider doing a needs assessment; a cost versus benefit analysis; assessing integration with existing systems; scalability; technical support and maintenance; training and workforce transition; regulatory compliance; and market size (Bernier, 2024). 

Reviewing business insurance policies for technology investment and failure coverage is also prudent as is establishing whether the business is covered for loses connected to cybersecurity.

The current state of AA is that its utility is in the eye of the beholder. Caution adoption of AA over time may be the wise choice for small to mid-sized growers. More widespread adoption in ten to twenty years from now, after the “kinks” in AA systems have been resolved may be the best ROI.

Read more from Dr. Brian Lynch here.

Sources

• Alam, H. (2023, Dec 11), A sustainable agricultural industry needs to start addressing mental health of farmers, says researcher, CBC.ca, 2023, https://www.cbc.ca/news/canada/edmonton/alberta-farmers-mental-health-anxiety-1.7055123
• Bernier, C. (2024, Feb 28), Robots in the greenhouse: Cultivating a new approach to agriculture, Cobot, https://howtorobot.com/expert-insight/robots-greenhouse-cultivating-new-approach-agriculture
• Connolly, B., (2024, Sept. 2). Canada’s food industry is vulnerable to cyberattacks, expert warns, https://www.ctvnews.ca/canada/canada-s-food-industry-is-vulnerable-to-cyberattacks-expert-warns-1.7022867
• Martin, D., & Rudolph, J., (2024, May 30) Manitoba Co-perator, https://www.manitobacooperator.ca/news-opinion/news/farm-risk-management-includes-cybersecurity/
• Nourish Labs https://nourishlabs.ca/
• Saba, R. (2024, June 16), CTV.ca, How AI could help farming become more efficient and sustainable, https://www.ctvnews.ca/business/how-ai-could-help-farming-become-more-efficient-and-sustainable-1.6928768
• TTControl Hydac International, How Europe’s farmers can leverage the power of AI, machine learning, and automation to feed a changing world, https://www.ttcontrol.com/farmers-leverage-ai-machine-learning-automation
• World Economic Forum (WEF) (2022, Jan 20), 3 ways autonomous farming is driving a new era of agriculture,  https://www.weforum.org/agenda/2022/01/autonomous-farming-tractors-agriculture/ 

• Tweet

• Foundation for growth: Root zone fundamentals
• Solar energy and the greenhouse climate