Survey Proposes Taxonomy for Federated Learning over Human-Body Communication

A recent survey paper available on arXiv (2605.24062) examines the convergence of human-body communication (HBC) and federated learning (FL) in the context of wearable edge intelligence. The authors highlight that HBC serves as an effective physical foundation for body-area networks by facilitating localized communication and alleviating radio link demands, while FL minimizes the centralization of raw data for physiological monitoring. Nonetheless, the connection between these two domains is limited: FL in wearables often overlooks the communication aspect, and HBC studies typically neglect learning and model-update traffic. The paper introduces a classification that differentiates between intra-body, body-hub, cross-user, and clinical-cloud FL implementations. It also addresses the challenge of body-channel-aware FL, which involves managing client selection, update compression, and aggregation through the body channel. The survey encompasses topics like wireless body-area networks, wearable FL, Internet-of-Bodies privacy, and optimization of edge intelligence.