Article · January 1, 2026

Real-Time Classification of Hand Movements from Surface Electromyography Signals Using Feature Engineering and an Ensemble of Machine-Learning Classifiers for Rehabilitation

Dilmurod Turimov, Dusmurod Kilichev, Norkobil Saydirasulov Saydirasulovic, Mukhamadiyev Abdinabi

Abstract

Myoelectric control of prostheses and biosignal-driven rehabilitation require handmovement recognition that is simultaneously accurate, computationally light, and fast enough forreal-time operation on embedded hardware. This paper presents an end-to-end methodology forclassifying object-interaction hand movements from surface electromyography (sEMG) signals.Four contributions are combined into a single deployable pipeline: (i) a real-time active-potentialonset/offset detection algorithm based on rectified-signal variance thresholding (threshold = μ +hσ); (ii) an optimal, low-dimensional feature vector obtained through feature engineering in thetime, frequency and time-frequency domains; (iii) a real-time algorithm for automatically buildingthe training set during signal acquisition; and (iv) an ensemble classifier that fuses k-NearestNeighbours, Support Vector Machine and Random Forest decisions with a confidence gate andtemporal confirmation. Two datasets were recorded - a single-channel BITalino dataset (20subjects, 4 movements) and a five-channel BTS FreeEMG dataset (42 subjects, 7 movements). Theproposed onset detector achieved the lowest mean error (15.45 samples) among six competingmethods.

Muscle activation and electromyography studiesContext-Aware Activity Recognition SystemsHand Gesture Recognition Systemshand motion recognitionsurface electromyographyreal-time classification

Real-Time Classification of Hand Movements from Surface Electromyography Signals Using Feature Engineering and an Ensemble of Machine-Learning Classifiers for Rehabilitation

Abstract

More by Norkobil Saydirasulov