Federated Deep Learning and TinyML Co-Design for Securing Resource-Constrained IoT and SCADA Networks: A PRISMA-Compliant Systematic Review, Taxonomy and Statistical Meta-Analysis
Abstract
Abstract As the Internet of Things (IoT) and industrial supervisory control and data acquisition (SCADA) systems expand, securing highly distributed, resource-constrained edge nodes against non-stationary, zero-day cyber threats has become a paramount challenge. While decentralized Federated Learning (FL) and localized TinyML co-design have emerged as promising paradigms to preserve data privacy and bypass wide-area network communication overhead, the existing literature remains highly fragmented. To resolve this, we present a comprehensive, PRISMA-compliant systematic review of federated deep learning and TinyML-IDS, synthesizing 237 peer-reviewed studies published between January 2020 and June 2026. First, we establish a multi-dimensional structural taxonomy that systematically categorizes localized neural cores (such as CNNs, LSTMs, Transformers, and Autoencoders), advanced federated optimization algorithms (including FedAvg, FedProx, FedNova, SCAFFOLD, and adaptive momentum-based variants), and cryptographic or statistical privacy defense frameworks (such as local differential privacy and secure multi-party aggregation). Secondly, we execute a rigorous, one-way Analysis of Variance (ANOVA) and post-hoc Tukey HSD analysis across the surveyed literature, mathematically validating that parallel Spatio-Temporal hybrid cores and attention-driven architectures yield a statistically superior mean detection accuracy (97.10%, p
How strong is the evidence?
Evidence Hub reads this as a meta-analysis. Its evidence level is 1 of 5 (Highest · pooled evidence). Pooled evidence from several studies sits at the top of the evidence pyramid.
Read the paper
Related studies
- Active Learning during Federated Learning for Object Detection
- Fairness-aware federated learning
- Decentralized Federated Learning
- Building a Personally Identifiable Information Recognizer in a Privacy Preserved Manner Using Automated Annotation and Federated Learning
- Preservation of Patient Level Privacy: Federated Classification and Calibration Models
- Efficient explainable federated learning
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.