Background: The Agri-food sector, along with other application domains, obtain high benefits from the monitoring capacity of Wireless Sensor Networks (WSN). This kind of applications should supply the users’ needs while integrated with different information systems. Such requirements could be seamlessly achieved in a model-driven approach through a complete and effective conceptual design phase. Current model-driven approaches allow to model WSN applications considering different communication and processing configurations. However, to the best of our knowledge, they do not allow for a complete and clear description of the node’s data and the aggregation operations applied to them.
Aims: The main aim of this thesis is to define a conceptual model for processing aggregation queries inside individual wireless sensor nodes of agriculture-oriented WSN. Modelling WSN data behaviour is relevant since it would allow to evaluate the capacity of an application for supplying the end-user needs, increase the cooperation capacity between the system’s designers, engineers, scientists and users, and define a transparent integration with different data-centric information systems.
Methods: To achieve this aim, this thesis was divided in four steps. In the first place, a systematic mapping with review allowed to describe the problem domain for an agricultural application. In the second place, a systematic literature review allowed to characterize the main wireless sensors used in agriculture. In the third place, a conceptual modelling process allowed to define an abstract representation of the wireless sensors’ data behaviour in UML, including the description of the data gathering, aggregation and delivering operations. Finally, in the fourth place, a thorough validation process with academics, case studies and CASE-Tool testing allowed to prove the feasibility of the proposed model.
Results: Through this process, we have defined a list with some of the most relevant challenges, characteristics and constraints in the design of agriculture-oriented WSN applications. From these features, we have built a conceptual meta-model in UML, which allows for the design of data querying in wireless sensor platforms, supporting the execution of temporal aggregation operations. Moreover, in this meta-model, we have achieved an explicit representation of the implicit separation between the data that is unavailable for the end-users and the data that is available and is used by the different information systems.
Conclusions: Our UML profile provides a standardized, complete and effective representation of the wireless sensors’ data behaviour from the user point-of-view, which could allow for the implementation of model-driven WSN applications that supply the end-user needs. Indeed, the models generated with our UML profile help visualise the ideal data behaviour in wireless-sensors-based systems, specifying their structure and operations, and allowing for the implementation of the real wireless sensors and their documentation. Moreover, since our UML profile provides different features for configuring the node operations, aggregating the gathered data, and checking the data quality, it allows increasing the user-perceived value of the WSN.