Novel approach of time series prediction using SEMG on hand gestures for prosthetic control

• Published in: Sadhana (Bangalore) Vol. 50; no. 3
• Main Authors: Sharma, Tanu, Sharma, K P
• Format: Journal Article
• Language: English
• Published: New Delhi Springer India 11.06.2025; Springer Nature B.V
• Subjects: Bioelectricity; Classification; Electromyography; Engineering; Muscles; Parameters; Prostheses; Signal processing; Time series; EMG signal; ARIMA model; signal processing; feature extraction; prosthetic
• ISSN: 0973-7677; 0256-2499; 0973-7677
• DOI: 10.1007/s12046-025-02759-1

To determine the purpose of a movement and assess the state of a muscle's function, one sort of bio-electrical signal i.e., electromyogram (EMG) signals can be used. There are several uses for EMG in motor control and neuromuscular physiology since it includes useful information about muscle activity. Electromyography (EMG) signal processing and categorization are crucial for controlling prosthetic arms and other applications. Recently, new methods for converting time series data to graphs and complicated networks with applications in biomedical (prosthetics) have been introduced. This ongoing study (i.e., the recording of seven hand activities from two independent muscle locations with ten subjects) suggests a time-series surface electromyogram (SEMG) signal pattern categorization approach for prosthesis control. For the prediction of time series data, the linear auto-regressive integrated moving average (ARIMA) model is investigated and evaluated. The regular activities performed could be distinguished greatly from the best classification results regarding R-squared and RMSE parameters. The findings from these data sets demonstrate that the suggested model has appreciable predicted parameter values and can be used for prosthetic applications.