Automated Channel Selection in High-Density sEMG for Improved Force Estimation

Abstract

Accurate and real-time estimation of force from surface electromyogram (EMG) signals enables a variety of applications. We developed and validated new approaches for selecting subsets of high-density (HD) EMG channels for improved and lower-dimensionality force estimation. First, a large dataset was recorded from a number of participants performing isometric contractions in different postures, while simultaneously recording HD-EMG channels and ground-truth force. The EMG signals were acquired from three linear surface electrode arrays, each with eight monopolar channels, and were placed on the long head and short head of the biceps brachii and brachioradialis. After data collection and pre-processing, fast orthogonal search (FOS) was employed for force estimation. To select a subset of channels, principal component analysis (PCA) in the frequency domain and a novel index called the power-correlation ratio (PCR), which maximizes the spectral power while minimizing similarity to other channels, were used. These approaches were compared to channel selection using time-domain PCA. We selected one, two, and three channels per muscle from the original seven differential channels to reduce the redundancy and correlation in the dataset. In the best case, we achieved an approximate improvement of 30% for force estimation while reducing the dimensionality by 57% for a subset of three channels.

Convolutional Neural Network Approach for Elbow Torque Estimation during Quasi-dynamic and Dynamic Contractions

“…Different approaches have been used to estimate torque, based on EMG signals [5], [6], [10], [11]. Some studies have used Hill's muscle model [11] in which an appropriate estimation of muscle physiological parameters is needed.…”

The Influence of the sEMG Amplitude Estimation Technique on the EMG–Force Relationship

“…Considering the fact that, for the tibialis dataset, the signals were extracted from an HD sEMG recording, the results presented here can be considered to be valid also in the case in which the sEMG–force relationship is exploited by this recording technology [ 3 , 21 , 25 ]. Estimating the amplitude of the different HD sEMG channel signals via an adaptive procedure can reasonably improve the outcomes of any processing algorithm that estimate force-related measures starting from the sEMG amplitude, yielding analogous performance differences with respect to the compact correlation measure that has been tested in this work using single-channel recordings.…”

“…In addition to the correlation measures, the root mean square error (RMSE) between the normalized time course of the force and the sEMG has been computed [ 21 ] as the RMS value of the difference between the two signals. Before calculating this parameter, the two segments were aligned to the delay corresponding to the maximum correlation, to compensate for physiological or instrumental time delays between the two quantities, such as the electromechanical delay [ 5 ].…”

Deep Multi-Scale Fusion of Convolutional Neural Networks for EMG-Based Movement Estimation

“…Many studies have focused on discrete motion control and used a classifier-based pattern recognition approach, where the EMG signals are used to discriminate discrete hand gestures [12], [13]. Regression models focus on continuous upper limb kinematics estimation [1]- [5], [14]- [16]. In the regression models, depending on the application, the estimated motor intent from the EMG during muscle contractions will be mapped to the generated force/torque [14]- [18], velocity [19],…”

“…Regression models focus on continuous upper limb kinematics estimation [1]- [5], [14]- [16]. In the regression models, depending on the application, the estimated motor intent from the EMG during muscle contractions will be mapped to the generated force/torque [14]- [18], velocity [19],…”