Abstract

As agricultural systems face mounting pressures from climate change, resource constraints, and growing food demands, digital technologies emerge as critical enablers of resilient and sustainable agri-food systems. Understanding how these technologies evolve and integrate with agricultural innovations is essential for shaping effective policies and investment strategies in the digital transformation of agriculture.
This paper proposes a novel methodology to identify and analyze smart farming innovation trajectories using comprehensive patent data from both the European Patent Office (EPO) and the United States Patent and Trademark Office (USPTO). Starting with the full set of patent applications filed between 1980 and 2025, the analysis leverages bibliometric data, including citations, to build complete network graphs for each patent office. Main path analysis using Search Path Node Pair (SPNP) methodology is applied to each network to identify the most influential innovation trajectories by tracing key knowledge flows through citation patterns. Smart farming technologies are identified through a hybrid approach combining WIPO digital patent classifications with OECD J-tag methodologies, capturing both recent innovations (AI, robotics, IoT) and foundational technologies (sensors, mobile communication, high-speed networks) that preceded more advanced digital applications.
The methodology extends beyond traditional patent classification approaches by incorporating climate change considerations through CPC code Y02 analysis and deploying a fine-tuned Large Language Model to extract technology characteristics and purposes from patent titles and abstracts. This LLM-based approach enables sophisticated clustering of technologies and identification of climate-focused innovations within smart farming trajectories. Patent trajectories are analyzed through their structural properties, temporal dynamics, technological composition, and sequencing patterns to understand how digital and agricultural innovations co-evolve over time.
Following established methodological frameworks, the analysis identifies leading countries in the technological race by calculating each nation's share of innovation and degree of specialization across different technology clusters. Cross-border knowledge flows are investigated through citation pattern analysis, revealing how agricultural and digital innovations diffuse internationally and identifying key knowledge hubs in the global smart farming ecosystem.
This comprehensive approach provides several contributions to understanding agricultural innovation systems in the digital age. First, it offers a quantitative framework for mapping the convergence of digital and agricultural technologies, revealing structural patterns in how smart farming innovations emerge and evolve. Second, the temporal analysis highlights waves of digital integration and shifting innovation dynamics, providing insights into technology adoption cycles and future innovation directions. Third, the climate change focus addresses whether agricultural innovations respond to environmental challenges. Finally, the international perspective reveals global innovation networks and competitive dynamics in agricultural technology development.
The findings will provide critical insights for policymakers, researchers, and industry stakeholders seeking to understand innovation patterns in agricultural digitalization. The methodology offers a scalable framework for monitoring technological convergence, identifying emerging innovation clusters, and tracking climate-responsive agricultural technologies. This evidence base can support informed decision-making in agricultural technology policy, research funding allocation, and investment strategies for building more resilient and sustainable agri-food systems in an era of digital transformation and climate uncertainty.