Federated Learning on Resource-Constrained IoT and TinyML Devices: review protocol and extraction data
Year: 2026
Type: article
Study design: Systematic Review
Abstract: Protocol (version 1.5, 6 November 2026) and supporting materials for a systematic literature review of federated learning (FL) on resource-constrained IoT and TinyML devices. The review identifies, classifies and synthesizes primary research on this topic, with emphasis on deployment evidence, resource optimization and feasibility, security and privacy overhead, benchmarking, and reproducibility. The protocol defines the research questions, the PICOC scope, the search strategy, the eligibility criteria, study selection, the device-tier classification, data extraction, quality assessment and the synthesis plan. Its amendment and deviation log records every change made after version 1.0 (19 August 2026), with its date and rationale. A copy with tracked changes shows what version 1.5 changed from version 1.4. The record also contains the search strings as run in each of the seven sources (searched 26–27 August 2026), the PICOC elements of each research question, and the completed data-extraction workbook for the 93 included studies (96 reports). In the workbook, each extracted value is linked to a verbatim quote and page in its source report. A README describes every file and workbook sheet.
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7225405095