Πλοήγηση ανά Συγγραφέας "Fragkopoulos, Markos-Vasileios"

Τώρα δείχνει 1 - 1 of 1
  • Μικρογραφία εικόνας
    Τεκμήριο
    End-to-End IoT platform concentrating in data and device diversity in Agriculture 4.0.
    (ΕΛ.ΜΕ.ΠΑ., Σχολή Μηχανικών (ΣΜΗΧ), ΠΜΣ Πληροφορική και Πολυμέσα, 2023-08-23) Fragkopoulos, Markos-Vasileios; Φραγκόπουλος, Μάρκος-Βασίλειος
    Smart farming can be defined as the application of supplementary technologies to help minimize waste and boost productivity. Agriculture sector involves a huge number of heterogeneous devices that are used for collecting, transferring, exchanging and processing data. Integration into a common infrastructure of diverse data from such devices is challenging due to compatibility issues. Data fusion, data transmission protocols, and serialization formats are essential components of agriculture IoT-based solutions as they enable seamless communication, data exchange, and interoperability. Despite the fact that IoT technologies are promptly evolving, some issues concerning the interoperability and the semantic annotation of heterogeneous data have to be handled within rural deployments that necessitate meeting certain requirements such as long range and coverage in areas with challenging terrains, where radio communications are difficult or not available. We have developed a platform that can deal with data and device diversity while supporting edge processing and dynamic context-based operation profiles for end nodes, by leveraging low energy consumption communication protocols and ultra simple end-to end deployment. On the edge there is an ARM-based single-board computer (SBC) Hybrid IoT node that is able to be adapted in any deployment. It’s RTOS is based on a distributed middleware that supports heterogeneity and offers flexibility on working both as an extreme edge dummy node, as edge computing node with processing capabilities, or as a Fog gateway able to communicate with subnetworks. Data transaction between the end nodes and the cloud is agnostic and is feasible through an application-level zero-copy binary serialization approach. For the data transfer LoRaWAN offers us flexibility due to its support for multiple spreading factors (SFs) and device classes. On the cloud core application, all IoT devices are distributed to virtual subnets that are handled by functionally independent resource managers. End devices can start working automatically once they are registered according to the operating scenario and system wide preferences. Decision making is done through adaptable computational models that could be fused, depending on the use case with third-party data in order to enrich the application context and improve the decision efficiency.