Converging Federated Learning, Edge Intelligence and Situational Awareness for Privacy-Preserving IoT Security: A Vision and Research Agenda

The rapid expansion of the Internet of Things (IoT) has created highly distributed environments in which devices continuously generate data relating to people, locations, activities, infrastructure, and operational processes. While this data can support intelligent security monitoring, conventional centralised approaches often require large volumes of potentially sensitive information to be transmitted to and processed by remote servers. This creates challenges relating to privacy, communication overhead, latency, data governance, and the security of distributed IoT infrastructures.

Federated Learning (FL) provides an alternative approach by allowing participating devices or edge nodes to train machine-learning models locally and share model updates rather than transferring raw data. When combined with Edge Intelligence, computation and decision-making can be moved closer to IoT data sources, potentially enabling more responsive security monitoring. However, machine learning alone does not provide a complete understanding of an IoT environment. Security decisions may also require situational awareness the ability to understand the current state of an environment, identify relevant events and anomalies, and place observations within their operational context.

This research therefore investigates the conceptual convergence of Federated Learning, Edge Intelligence, Situational Awareness, and privacy-preserving IoT security. The central research question is how these technologies can be integrated into a coherent architecture capable of supporting distributed security intelligence while reducing unnecessary exposure of sensitive IoT data.

The project will examine the relationships between local intelligence, collaborative model learning, contextual information, privacy protection, threat detection, and security decision-making. Particular attention will be given to challenges arising from heterogeneous IoT devices, non-IID and imbalanced data, resource-constrained edge environments, communication limitations, adversarial participants, model poisoning, privacy leakage, and the need to evaluate security performance beyond conventional aggregate accuracy measures.

Rather than treating the convergence of these technologies as an established solution, the study will identify the assumptions, research gaps, unresolved technical challenges, and evaluation requirements that must be addressed before such systems can be reliably developed and deployed.

The intended outcome is a structured research agenda and conceptual foundation that can guide subsequent empirical studies, prototype development, benchmark construction, and interdisciplinary research. The project is particularly suited to collaboration between researchers working in Federated Learning, Edge AI, IoT, Cybersecurity, Privacy-Preserving Machine Learning, Situational Awareness, Distributed Systems, and Intelligent Computing.

Researchers joining the project may contribute through literature synthesis, conceptual modelling, mathematical formulation, system architecture, dataset and benchmark development, algorithm design, experimental evaluation, cybersecurity analysis, privacy assessment, or prototype implementation.