Autoencoder-based transceivers for multiple access human body communication networks

Ali, Abdelhay, Abdelrahman, Amr N., Celik, Abdulkadir and Eltawil, Ahmed M. (2025) Autoencoder-based transceivers for multiple access human body communication networks. IEEE Transactions on Circuits and Systems I: Regular Papers. (doi:10.1109/TCSI.2025.3582415).

Abstract

The Internet of Bodies (IoB) represents a transformative technological innovation that merges the bio-physical and digital realms through networks of intelligent devices positioned in, on, and around the human body. Human Body Communication (HBC) offers a promising method for enabling IoB networks, using the human body as a communication channel for multiple wearable nodes. Despite the prevalence of HBC peer-to-peer communication methods in the literature, the challenge of implementing multiple access techniques for HBC at the physical layer remains largely unexplored. In this paper, we propose a new multiple access HBC (MA-HBC) system that leverages autoencoders to design and implement low-power and efficient transceivers sharing a common channel. The proposed MA-HBC system consists of multiple autoencoder-based transceivers trained jointly to optimize overall network performance. It supports various data rates ranging from 164 kbps to 5.25 Mbps, making it suitable for a wide range of IoB applications. To validate the design, a prototype implementation is presented. Additionally, to ensure suitability for wearable devices, a low-power hardware transceiver architecture is provided with an estimated energy efficiency of 105 pJ/b when implemented using TSMC 65nm LP technology. The results show that the proposed MA-HBC system outperforms traditional IEEE 802.15.6 based transceivers for two users with time Sharing, achieving a 3.9 dB improvement in Signal-to-Noise Ratio (SNR) at a block error rate of 10⁻² .

This record has no associated files available for download.

Contributors

Author: Abdulkadir Celik

Download statistics