Khan, Sagheer and Anwar, Usman and Khurshid, Kiran and Ullah, Rahmat and Arslan, Tughrul and Ahmed, Moataz (2026) AI-Driven EMG Monitoring and Decision Support Framework. IEEE Sensors Journal. p. 1. DOI https://doi.org/10.1109/jsen.2026.3659791
Khan, Sagheer and Anwar, Usman and Khurshid, Kiran and Ullah, Rahmat and Arslan, Tughrul and Ahmed, Moataz (2026) AI-Driven EMG Monitoring and Decision Support Framework. IEEE Sensors Journal. p. 1. DOI https://doi.org/10.1109/jsen.2026.3659791
Khan, Sagheer and Anwar, Usman and Khurshid, Kiran and Ullah, Rahmat and Arslan, Tughrul and Ahmed, Moataz (2026) AI-Driven EMG Monitoring and Decision Support Framework. IEEE Sensors Journal. p. 1. DOI https://doi.org/10.1109/jsen.2026.3659791
Abstract
Digital Twin (DT) technology, a core pillar of Healthcare 4.0 (H4.0), enables intelligent, non-invasive, and personalized patient monitoring. This research presents a pilot AI-enabled Electromyography (EMG)-driven DT framework for muscular activity assessments. The EMG data enables monitoring of muscle engagements and provides a comprehensive representation of physiological states. The raw EMG data, consisting of 7 activities, i.e., Sitting, Standing, Walking, Relax, Stress Ball, Hand at Rest, and Fist, is subjected to denoising techniques of mean removal, smoothing, and digital filtering. Within the DT model, AI serves as the intelligence core that transforms these denoised signals into relevant digital states. Supervised and semi-supervised classifiers act as inference engines, continuously refining the DT as new data is incorporated, allowing it to evolve in synchrony with the patient’s condition. The decision support using ML and DL is employed for EMG classification, utilizing statistical features and autonomous feature extraction methodologies of AutoEncoder (AE) and Stacked AutoEncoder (SAE). The feature data is enriched and enlarged through Gaussian noise feature data augmentation for both feature extraction approaches. The Fine KNN algorithm provides classification accuracy of 94.6% and 91.6%. However, the autonomous feature extraction through the SAE (32-16-32) with Medium KNN provides an overall accuracy of 96.4% and 93.3%. The promising results validate the effectiveness of the proposed framework as a dynamic, AI-driven DT system for prospective holistic patient multi-physiological monitoring and decision support.
| Item Type: | Article |
|---|---|
| Uncontrolled Keywords: | Digital Twin; Electromyography; Classification; Multi-Physiological; Signal Processing; Machine Learning; Deep Learning |
| Subjects: | Z Bibliography. Library Science. Information Resources > ZR Rights Retention |
| Divisions: | Faculty of Science and Health Faculty of Science and Health > Computer Science and Electronic Engineering, School of |
| SWORD Depositor: | Unnamed user with email elements@essex.ac.uk |
| Depositing User: | Unnamed user with email elements@essex.ac.uk |
| Date Deposited: | 09 Feb 2026 13:57 |
| Last Modified: | 09 Feb 2026 14:08 |
| URI: | http://repository.essex.ac.uk/id/eprint/42776 |
Available files
Filename: AI-Driven_EMG_Monitoring_and_Decision_Support_Framework.pdf
Licence: Creative Commons: Attribution 4.0