New machine learning advancements will help optimize harvest equipment and improve on-farm efficiencies

Improved efficiencies at harvest have the potential to ripple through the value chain, leading to benefits from field to fork. In an effort to improve modern harvest equipment to help make operators more efficient and reduce labour demand, MacDon is partnering with the University of Manitoba on a new Protein Industries Canada project that will accelerate the development and deployment of automation technologies for Canada’s agriculture sector.

“AI and machine learning have the potential to be a foundational element in Canada’s Make It Here opportunity, helping our country strengthen its economy and domestic food supply chain,” Protein Industries Canada CEO Tyler Groeneveld said. “By expanding the on-farm use of these technologies, our food production and value-added agriculture sector can become more efficient, more sustainable and more productive, leading to a stronger Canadian economy and new opportunity in the global food and ingredient market.”

Throughout the project, MacDon and the University of Manitoba will create synthetic datasets that replicate field conditions, enabling faster model development, iteration and testing without relying on seasonal in-field data collection. This use of synthetic data will help the partners to reduce traditional model development timelines, giving them the opportunity to put it into Canadian farmers’ hands sooner.

This, combined with the technology’s potential to train machine learning (ML) models in farm machinery, will help improve on-farm productivity, reduce operator burden and improve operator effectiveness—leading to a more competitive Canadian food production and value-added agriculture sector.

“ML has many potential applications within agriculture, however a major challenge with ML technology is capturing sufficient data to address the variation between farms. Through this project, MacDon will be improving upon its existing methods of applying ML technology, while finding new uses for the technology and accelerating its application within our product lines. MacDon will work directly with producers and custom harvesters in the most challenging real-world conditions to assess the effectiveness of these approaches and maximizing the value for our customers,” said Ryan Georgison, MacDon’s R&D Engineering Manager.

Through improved AI technology and the development of new opportunities for synthetic data and machine learning capabilities, partners MacDon and the University of Manitoba are helping strengthen Canada’s innovation ecosystem and domestic food supply chain. A total of $400,000 has been committed to the project, with Protein Industries Canada committing $180,000 and the partners together committing the remainder. Protein Industries Canada’s artificial intelligence stream is funded as part of the Pan-Canadian Artificial Intelligence Strategy (PCAIS), through which the Government of Canada is investing in efforts to drive the adoption of artificial intelligence across Canada’s economy and society.