Hand Grasps Recognition Using Single-Channel Transient-State Surface Electromyography
Nasim Alnuman,Muhannad Abdullah,Merna Haddad,Jumana Ma’touq,Jumana Ma'touq
DOI: https://doi.org/10.4015/s1016237224500236
2024-06-30
Biomedical Engineering Applications Basis and Communications
Abstract:Biomedical Engineering: Applications, Basis and Communications, Ahead of Print. The need to perform large hand-grasping strategies in daily life has promoted the development of robotic hands and prosthetic hands that mimic natural ones. This work presents a system for the detection of 10 basic hand grasps in healthy subjects based on single-channel transient-state Surface Electromyography (sEMG) signals. The studied grasps are large diameter, ring, knife index finger, writing tripod, power sphere, precision sphere, prismatic pinch, lateral, extension type, and parallel extension grasp. Signals were collected from 10 healthy volunteers using three sEMG electrodes connected to an Electromyography (EMG) shield using MATLAB as data acquisition software. The signals were classified using three different classifiers, i.e. [math]-Nearest Neighbors ([math]-NN), Diagonal Linear Discriminant Analysis (dl-DA), and Random Forest (RF). The used features were mean absolute value, integrated absolute value, variance, root mean square value, waveform length, Willison amplitude, slope sign change, zero crossing, autoregressive coefficients, and mean frequency. The results show that good classification accuracy is achieved when classifying a reduced number of grasps using a single-channel transient-state sEMG. The average accuracy for the different classifiers varied from 41.5% to 53.3% when classifying 10 grasps and from 78.4% to 91.9% when classifying five grasps using two features as a trainer. It is found that transient-state sEMG signals are efficient in the detection of different hand grasps, reducing the time to generate control signals and the number of channels.