Files
Abstract
The shift from conventional to smart farming has become urgent due to the growing global population and resource scarcity exacerbated by climate change. Smart farming, driven by the transformative potential of the Internet of Things (IoT), offers a solution. Through seamless integration of advanced IoT technologies – sensors, data analytics, and automation – farmers optimize productivity while conserving vital resources. Instant data from IoT sensors helps them make smarter decisions, adjusting plans as conditions shift. This clarity ensures efficient operations and resource management, ultimately benefiting both farmers and the environment.
As the smart farming ecosystem grows, robust access control becomes crucial, ensuring secure data and system management against cyber threats from authorized and unauthorized interferences. Although there has been good progress in access control solutions tailored to other smart ecosystems, smart farm-specific access control solutions remain a challenging area that requires further research and development to effectively address the unique requirements and complexities of farm environments.
This dissertation explores the unique characteristics of smart farms and highlights the importance of proposing effective and efficient access controls tailored specifically for them. Toward creating an effective and efficient solution, we performed a multifaceted exploration into the security scenarios arising from the diverse nature of smart farms. Additionally, we explore fundamental questions essential for identifying an optimal access control mechanism tailored to the unique characteristics of farms. Moreover, We introduce a semantically enriched access control architecture designed specifically for smart farms, with the primary goal of simplifying the process of obtaining additional attributes necessary for access request decisions and access control policies administration tasks achieved through the utilization of a meticulously crafted smart farm ontology. We further extend the architecture to a content-driven one that incorporates an edge machine learning classifier to make informed decisions regarding access requests for images based on their visual content.
We evaluated deploying the proposed architecture variations on the edge area to better suit the smart farm nature. Our experiments show that the proposed variations in architecture markedly improve access control decision-making within smart farming environments. Notably, these enhancements cater to the constraints of edge-limited devices, ensuring swift and efficient performance.
As the smart farming ecosystem grows, robust access control becomes crucial, ensuring secure data and system management against cyber threats from authorized and unauthorized interferences. Although there has been good progress in access control solutions tailored to other smart ecosystems, smart farm-specific access control solutions remain a challenging area that requires further research and development to effectively address the unique requirements and complexities of farm environments.
This dissertation explores the unique characteristics of smart farms and highlights the importance of proposing effective and efficient access controls tailored specifically for them. Toward creating an effective and efficient solution, we performed a multifaceted exploration into the security scenarios arising from the diverse nature of smart farms. Additionally, we explore fundamental questions essential for identifying an optimal access control mechanism tailored to the unique characteristics of farms. Moreover, We introduce a semantically enriched access control architecture designed specifically for smart farms, with the primary goal of simplifying the process of obtaining additional attributes necessary for access request decisions and access control policies administration tasks achieved through the utilization of a meticulously crafted smart farm ontology. We further extend the architecture to a content-driven one that incorporates an edge machine learning classifier to make informed decisions regarding access requests for images based on their visual content.
We evaluated deploying the proposed architecture variations on the edge area to better suit the smart farm nature. Our experiments show that the proposed variations in architecture markedly improve access control decision-making within smart farming environments. Notably, these enhancements cater to the constraints of edge-limited devices, ensuring swift and efficient performance.