A Novel Method for Hand Movement Recognition Based on Wavelet Packet Transform and Principal Component Analysis with...

Area of Science:

• Biomedical Engineering
• Human-Computer Interaction
• Signal Processing

Background:

• Surface electromyogram (sEMG) signals offer rich information for human movement and natural human-computer interaction.
• Effective component extraction from sEMG signals is crucial for improving hand motion classification accuracy.
• Current methods face challenges in optimizing feature extraction for enhanced classification.

Purpose of the Study:

• To propose a novel method for classifying six hand motions using sEMG signals.
• To enhance the accuracy of hand motion classification through advanced signal processing techniques.
• To evaluate the effectiveness of the proposed method in real-world applications.

Main Methods:

• Utilized Wavelet Packet Transform (WPT) to decompose sEMG signals into sub-band signals.
• Evaluated different wavelet packet basis functions for efficient intrinsic component extraction.
• Applied Principal Component Analysis (PCA) for feature space dimensionality reduction.
• Analyzed the impact of sEMG signal variability and window size on classification performance.

Main Results:

• Achieved an average classification accuracy of 96.03% on the sEMG for Basic Hand Movements Data Set.
• Demonstrated superior classification performance compared to existing research methods.
• Identified optimal wavelet packet basis functions and feature space dimensions for PCA.

Conclusions:

• The proposed WPT-PCA method effectively classifies hand motions from sEMG signals with high accuracy.
• This research provides a robust approach for human-computer interaction applications.
• Potential applications include exoskeleton robots, rehabilitation training, and intelligent prosthetics.