Distributed Intelligence in Smart Agriculture Through Cognitive Tree Modeling

Smart Agriculture is increasingly adopting intelligent sensing and data-driven techniques to improve sustainability, productivity, and crop efficiency. In this context, the ability to continuously monitor and interpret the physiological status of plants, particularly water stress in woody species, is a key challenge. While existing solutions largely focus on technological and architectural aspects—such as IoT platforms, edge/cloud infrastructures, and AI-based analytics—there is a lack of explicit methodological approaches for systematically modeling such complex systems and their data properties. This paper proposes the concept of a Cognitive Tree, a domain-specific specialization of the Cognitive Environment paradigm tailored to precision agriculture. A Cognitive Tree is defined as a distributed intelligent system, where multiple smart functionalities—spanning sensing, monitoring, recognition, prediction, and control—cooperate to enable plant-level awareness and adaptive decision-making. In this perspective, intelligence is not centralized but results from the coordinated interaction of distributed functionalities, each contributing to the overall system behavior. To support the systematic design of such systems, the Smart Environment Meta-model (SEM) is adopted as a unifying methodological framework, enabling a structured representation of functional components, data characteristics, and their interactions. The proposed approach is validated through the design and deployment of a real Cognitive Tree prototype, enabling continuous assessment of tree water stress. Overall, this work demonstrates how a metamodel-driven approach can support the design of distributed and cooperative intelligent systems in smart agriculture, promoting modularity, extensibility, and system-level intelligence grounded in the interaction of multiple functionalities.