Abstract
Modern agricultural IoT systems generate large volumes of heterogeneous data; however, the reuse of this data and cross-system integration remain challenging due to the lack of a unified semantic representation and mechanisms for controlling information usage. This paper proposes a middleware layer architecture for preliminary semantic processing of agricultural data in a distributed edge environment. As part of the research, an experimental proof-of-concept prototype was designed and partially implemented, including a containerized deployment environment, an automatic semantic mapping mechanism, and the logging of system data processing events. Individual parameters from the open RHoMIS dataset, mapped to terms from the AgrO and ENVO agricultural ontolo-gies, were used for testing. The results showed that the proposed rule-based approach allows for the automatic preliminary FAIR adaptation of agricultural data at relatively low computational cost. Preliminary experiments also confirmed the middleware’s ability to function under the resource constraints of edge devices.