Validation of a selective ensemble-based classification scheme for myoelectric control using a three-dimensional...

Area of Science:

• Biomedical Engineering
• Rehabilitation Engineering
• Human-Computer Interaction

Background:

• Effective control of powered upper limb prostheses requires not only executing intended movements but also preventing unintended ones.
• Existing pattern recognition control schemes may struggle with muscle signals not corresponding to desired actions, leading to suboptimal performance.
• Evaluating real-time myoelectric control performance traditionally relies on virtual limb environments, which may not fully capture real-world complexities.

Purpose of the Study:

• To introduce and evaluate a novel selective multiclass one-versus-one classification scheme for powered upper limb prosthesis control.
• To propose and validate a 3-D Fitts' Law test as a robust alternative to virtual limb environments for assessing myoelectric control.
• To compare the performance of the selective classification scheme against a state-of-the-art linear discriminant analysis (LDA) based scheme using the proposed Fitts' Law test.

Main Methods:

• Implementation of a selective multiclass one-versus-one classification strategy for myoelectric signal processing.
• Development and application of a 3-D Fitts' Law task to evaluate real-time prosthesis control.
• Comparative analysis of the selective scheme and an LDA-based scheme using Fitts' Law metrics and additional quality-of-control indicators.

Main Results:

• The 3-D Fitts' Law framework demonstrated high linearity (R(2) > 0.936) for both control schemes, validating its use.
• The selective classification scheme achieved significantly higher efficiency and completion rates compared to the LDA scheme.
• The selective scheme resulted in significantly lower overshoot and stopping distances, indicating superior control precision.

Conclusions:

• The selective multiclass one-versus-one classification approach effectively rejects unintended motions in powered upper limb prostheses.
• The 3-D Fitts' Law test provides a reliable and valid method for evaluating real-time myoelectric control performance.
• The proposed selective control strategy offers significant improvements in prosthesis control quality and user performance.