Design and Experimental Realization of Wearable UWB Antennas for Machine Learning–Driven Human Posture and Activity Recognition
International Journal of Antennas and Propagation · 2026-01 · 2026 (1)
DOI: 10.1155/ijap/7745584 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Antennas and Propagation, online erschienen am 25.06.2026. Verfasst von Bhawna Tiwari und 3 weiteren Autor:innen. Die Arbeit nennt 50 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Combining wearable sensing technologies with intelligent classification algorithms opens new possibilities for tracking human activities and postures, with applications in healthcare, rehabilitation, and human–computer interaction. In the present scenario of wearable‐based human activity monitoring, body‐worn equipment can facilitate real‐time observation by integrating sensing systems onto the human body. The extensive focus of this research is to develop and implement an on‐body ultra‐wideband antenna based on FR4 and jeans substrate tailored for activity recognition applications and to demonstrate its feasibility to be used for a human body posture recognition and classification system. This research work reports the development of compact body‐worn UWB antennas integrated with intelligent machine learning algorithms to achieve reliable and high‐accuracy human posture and activity recognition. The antennas are optimized for stable on‐body performance, ensuring reliable signal acquisition during dynamic movements. Key scattering parameters (S 11 and S 21 ) are used as features for classification. The compact wearable transmitting and receiving antennas are designed and implemented to capture datasets of on‐body antenna characteristics, including reflection and transmission parameters, voltage standing wave ratio (VSWR), reflected power, and mismatch loss associated with various human body postures, performed by different subjects. The comparative performance exploration of various machine learning classifiers, such as random forests (RFs), decision trees (DTs), support vector machines (SVMs), Gaussian Naïve Bayes (GNB), K‐nearest neighbors (KNNs), extreme gradient boosting (XGB), and neural networks for processing, analyzing, and classification purposes, is performed based on datasets observed for different body postures. Enhanced classification performance is achieved with the GNB and DT classification algorithms. Experimental results showcase high classification accuracy and clear distinguishability across posture states. These findings establish the potential of UWB wearable antennas combined with AI‐driven models for intelligent healthcare monitoring and activity‐aware applications.
Zitieren
APA
Tiwari, B., Gupta, S. H., Balyan, V., & Ojha, M. K. (2026). Design and Experimental Realization of Wearable UWB Antennas for Machine Learning–Driven Human Posture and Activity Recognition. International Journal of Antennas and Propagation, 2026(1). https://doi.org/10.1155/ijap/7745584
Vancouver
Tiwari B, Gupta SH, Balyan V, Ojha MK. Design and Experimental Realization of Wearable UWB Antennas for Machine Learning–Driven Human Posture and Activity Recognition. International Journal of Antennas and Propagation. 2026;2026(1). doi:10.1155/ijap/7745584
BibTeX
@article{Tiwari2026,
author = {Tiwari, Bhawna and Gupta, Sindhu Hak and Balyan, Vipin and Ojha, Mukesh Kumar},
title = {Design and Experimental Realization of Wearable UWB Antennas for Machine Learning–Driven Human Posture and Activity Recognition},
journal = {International Journal of Antennas and Propagation},
year = {2026},
volume = {2026},
number = {1},
doi = {10.1155/ijap/7745584}
}
EUVIMED