Hand movement recognition from sEMG signals using Fourier decomposition method

Abstract

Surface electromyogram (sEMG) provides a non-invasive way to collect EMG signals. The sEMG signals acquired from the muscles of the forearm can be used to recognize the hand grasps and gestures. In this work, an automatic recognition algorithm to identify hand movements using sEMG signals has been proposed. The signals are decomposed into Fourier intrinsic band functions (FIBFs) using the Fourier decomposition method (FDM). The features like entropy, kurtosis, and L1 norm are computed for each FIBF. Statistically relevant features are determined using the Kruskal Wallis test and used to train machine learning-based classifiers like support vector machine, k-nearest neighbor, ensemble bagged trees, and ensemble subspace discriminant. Two publicly available datasets are used to test the efficacy of the proposed algorithm. With an average accuracy of 99.49% on the UCI dataset and 93.53% on NinaPro DB5, the proposed method performs superior than the state-of-the-art algorithms. The performance of the proposed algorithm has also been analyzed in the presence of noise. The proposed method is based on Fourier theory, which makes it suitable for real-time implementation due to low computational complexity. It would help in the design of efficient and easy-to-use prosthetic hands. © 2021 Nalecz Institute of Biocybernetics and Biomedical Engineering of the Polish Academy of Sciences