Robust Multi-Frequency Posture Detection based on Purely Passive Magneto-Inductive Tags

Authors

Henry Schulten and Armin Wittneben

Reference

IEEE International Conference on Communications (ICC), Seoul, South Korea, pp. 5, 2022-05-16.

DOI: 10.1109/ICC45855.2022.9839034

[BibTeX, LaTeX, and HTML Reference]

Abstract

Recent works have demonstrated the suitability of magnetic induction for future wireless and wearable posture detection systems. They showed that such systems can work with a low complexity by relying on purely passive resonant agent coils on all limbs while multiple anchor coils are centralized on the torso with their input impedance measured at a single frequency. In this work, we experimentally study the practical robustness of such a low-cost, low-complexity posture detection approach by introducing posture variations and unintentional, random displacements of the coils. We find that the prior supervised single-frequency classifiers degrade substantially with these additional disturbances unless they are also present in the training data. Motivated by an in-depth consideration of the degradation mechanism, we propose to measure the complex anchor impedance at multiple frequencies and provide numerical experimental evidence of the improved robustness. To facilitate system design, we further quantify the trade-off between classification accuracy, center frequency and bandwidth. We conclude that a recommendable operating point is a carrier frequency of 510kHz with a measurement bandwidth of 40kHz. With these specifications, a classification accuracy of 90 % is retained even in the case of reduced training data, that does not comprise the additional disturbances of the test data. In order to mitigate the inherent complexity increase of the multi-frequency operation, we further propose to use the magnitude of the anchor impedances as features for the deployed classifiers. This simplification results in a negligible accuracy degradation of less than 3 %.


Download this document:

 

Copyright Notice: © 2022 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.

This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright. In most cases, these works may not be reposted without the explicit permission of the copyright holder.