Recognition of Unseen Combined Motions via Convex Combination-based EMG Pattern Synthesis for Myoelectric Control

TL;DR

This work tackles the data-collection bottleneck in EMG-based motion recognition by generating synthetic EMG patterns for unseen combined motions via convex combinations of basic motion patterns. A neural-network classifier is trained on basic motions plus synthetic combined-data, with synthesis performed at an input or hidden layer, guided by a symmetric Dirichlet sampling of mixing coefficients $\lambda_k$ and a composite loss. A similarity-based evaluation using KL-divergence with basis vectors enables recognition of combined motions, even when only basic-motion data are available for training. Experiments with eight participants show that input-layer synthesis yields the best performance, achieving improved recognition of unseen combined motions, though linear synthesis cannot fully replicate complex muscular co-contraction patterns; future work proposes non-linear mappings and cross-subject transfer to further boost accuracy.

Abstract

Electromyogram (EMG) signals recorded from the skin surface enable intuitive control of assistive devices such as prosthetic limbs. However, in EMG-based motion recognition, collecting comprehensive training data for all target motions remains challenging, particularly for complex combined motions. This paper proposes a method to efficiently recognize combined motions using synthetic EMG data generated through convex combinations of basic motion patterns. Instead of measuring all possible combined motions, the proposed method utilizes measured basic motion data along with synthetically combined motion data for training. This approach expands the range of recognizable combined motions while minimizing the required training data collection. We evaluated the effectiveness of the proposed method through an upper limb motion classification experiment with eight subjects. The experimental results demonstrated that the proposed method improved the classification accuracy for unseen combined motions by approximately 17%.

Figures (8)

• Figure 1: Schematic overview of the proposed method for generating synthetic EMG patterns of combined motions using convex combinations of basic motion patterns at different neural network layers
• Figure 2: Illustration of similarity calculation between predicted probabilities and basis probability vectors for motion classification
• Figure 3: Examples of basic motions (S1, S4) and combined motions (C11)
• Figure 4: Comparison of actual and synthetic EMG patterns using radar plots for six combined motions (C3, C4, C7, C8, C9, and C12)
• Figure 5: Examples of principal component analysis visualization comparing actual and synthetic EMG patterns at the (a) input layer and (b) first hidden layer for three motion classes (S2, S5, and C2).