Cascade of one-class classifier ensemble and dynamic naive Bayes classifier applied to the myoelectric-based upper limb prosthesis control with contaminated channels detection

TL;DR

This paper tackles the problem of contaminated sEMG channels in upper-limb prosthesis control by proposing a two-stage cascade: an ensemble of one-class classifiers detects channel contamination, and a dynamic Naive Bayes classifier uses this information to infer movement with one-shot training. The key contribution is the dynamic integration of contamination signals, expressed through a posterior form $d_j(x,r)=c \cdot p_j \cdot \prod_{l=1}^{L} (\prod_{i=1}^{d_l} P(x_l^{(i)}|j))^{r_l}$, which allows weighting or removal of channel features without retraining per object. Experimental results across various SNRs show that soft-contamination weighting within the dynamic NB framework (NBS) and its hard counterpart (NBH) outperform baselines and an EC ensemble, especially under higher contamination. The findings suggest a practical, low-latency strategy for robust sEMG-based prosthesis control in noisy real-world environments and point to further exploration of dynamic NB in this domain.

Abstract

Modern upper limb bioprostheses are typically controlled by sEMG signals using a pattern recognition scheme in the control process. Unfortunately, the sEMG signal is very susceptible to contamination that deteriorates the quality of the control system and reduces the usefulness of the prosthesis in the patient's everyday life. In the paper, the authors propose a new recognition system intended for sEMG-based control of the hand prosthesis with detection of contaminated sEMG signals. The originality of the proposed solution lies in the co-operation of two recognition systems working in a cascade structure: (1) an ensemble of one-class classifiers used to recognise contaminated signals and (2) a naive Bayes classifier (NBC) which recognises the patient's intentions using the information about contaminations produced by the ensemble. Although in the proposed approach, the NBC model is changed dynamically, due to the multiplicative form of the classification functions, training can be performed in a one-shot procedure. Experimental studies were conducted using real sEMG signals. The results obtained confirm the hypothesis that the use of the one-class classifier ensemble and the dynamic NBC model leads to improved classification quality.

Figures (2)

• Figure 1: Classification quality for NBG classifier -- average ranks.
• Figure 2: Classification quality for NBGMT classifier -- average ranks.