<![CDATA[The rapid growth of Internet of Things (IoT) ecosystems has increased security and privacy risks due to the large volume of sensitive and distributed data generated by connected devices. Traditional machine learning approaches for IoT security rely on centralized data aggregation, which raises privacy concerns, scalability limitations, and regulatory challenges. This study presents a systematic literature review of privacy-enhanced federated learning (PEFL) for secure IoT applications. The review analyzes federated learning architectures, optimization strategies, privacy-preserving mechanisms, and cybersecurity use cases in heterogeneous IoT environments. Findings indicate that federated learning improves privacy by keeping data localized at edge devices, while techniques such as differential privacy and secure aggregation enhance confidentiality. However, these methods introduce trade-offs in accuracy, communication efficiency, and computational overhead. Hierarchical and adaptive FL architectures improve scalability and robustness, particularly for intrusion detection systems. Persistent challenges include non-IID data, device heterogeneity, limited edge resources, and the lack of standardized benchmarks and real-world deployments. PEFL offers a promising framework for securing IoT systems, though further research is needed to address deployment and performance constraints in practical environments.]]>
Copyrights © 2026
