A dual ensemble classifier used to recognise contaminated multi-channel EMG and MMG signals in the control of upper limb bioprosthesis

TL;DR

This work tackles robust myopotential pattern recognition for upper-limb bioprostheses in the presence of multimodal, multichannel biosignals subject to contamination. It introduces a dual ensemble architecture: a one-class ensemble per channel to detect contaminants, and a second multiclassifier ensemble for movement classification with dynamic selection that excludes classifiers using contaminated channels, implemented via a rule $S$ and majority voting. Features are extracted per channel using discrete wavelet transform with $db6$ at three levels, and Random Forests with 30 trees serve as base classifiers, complemented by a tunable one-class SVM detector. Experiments on an able-bodied subject with simulated amputation across eight imagined movements show enhanced robustness under low $SNR$ when aggregating across multiple values of $K$, highlighting the potential of contamination-aware dynamic ensembles for real-world myopotential control.

Abstract

Myopotential pattern recognition to decode the intent of the user is the most advanced approach to controlling a powered bioprosthesis. Unfortunately, many factors make this a difficult problem and achieving acceptable recognition quality in real-word conditions is a serious challenge. The aim of the paper is to develop a recognition system that will mitigate factors related to multimodality and multichannel recording of biosignals and their high susceptibility to contamination. The proposed method involves the use of two co-operating multiclassifier systems. The first system is composed of one-class classifiers related to individual electromyographic (EMG) and mechanomyographic (MMG) biosignal recording channels, and its task is to recognise contaminated channels. The role of the second system is to recognise the class of movement resulting from the patient's intention. The ensemble system consists of base classifiers using the representation (extracted features) of biosignals from different channels. The system uses a dynamic selection mechanism, eliminating those base classifiers that are associated with biosignal channels that are recognised by the one-class ensemble system as being contaminated. Experimental studies were conducted using signals from an able-bodied person with simulation of amputation. The results obtained allow us to reject the null hypothesis that the application of the dual ensemble foes not lead to improved classification quality.

Figures (5)

• Figure 1: Illustration of recording EMG and MMG signals.
• Figure 2: The impact of changing $K$, signal with noise -- ranks.
• Figure 3: The impact of changing $K$, signal without noise -- ranks.
• Figure 4: The impact of changing SNR, $K=7$ -- boxplots.
• Figure 5: The impact of changing SNR, $K \in \{2, 3, 5\}$ -- boxplots.