Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

MA: Fachverband Magnetismus

MA 49: Magnetic Imaging, Information Technology, and Sensors

MA 49.5: Vortrag

Donnerstag, 12. März 2026, 16:00–16:15, POT/0112

Deep learning enabled wearable magnetoelectronics — •Guannan Mu, Rui Xu, Proloy T. Das, Jan Schmidtpeter, Lin Guo, Oleksandr Pylypovskyi, Andreas Knüpfer, Rico Illing, Olha Bezsmertna, and Denys Makarov — Helmholtz-Zentrum Dresden-Rossendorf e.V., 01328 Dresden, Germany

A wide variety of magnetic field sensors is already integrated into human-machine interfaces in wearable electronics [1,2]. However, the currently available interfaces are limited in the number of recognizable gestures and operation distance. To overcome these limitations, we integrate deep learning into two magnetic interaction platforms. First, we employ a flexible magnetoresistive sensor for user-definable temporal multipattern classification. Furthermore, we realize a compact LSTM-enabled wearable magnetic-interaction system integrated into a wrist-worn platform that incorporates a planar Hall magnetoresistive sensor, enabling a single sensor to recognize multiple predefined gestures and achieving a high classification accuracy of 99.4 % at a long range interaction distance of 12 cm, highlighting the potential of deep learning-enhanced magnetic sensing to expand functionality and enable long-range magnetic interaction recognition in smart wearable interfaces. [1] R. Xu et al., ACS Nano 19, 21891 (2025) [2] O. Bezsmertna et al., Adv. Funct. Mater. 2502947 (2025)

Keywords: Deep Learning; wearable electronics; magnetic interaction; multipattern; long range interaction