sEMG-Based Drawing Trace Reconstruction: A Novel Hybrid Algorithm Fusing Gene Expression Programming into Kalman...

Area of Science:

• Biomedical Engineering
• Signal Processing
• Machine Learning

Background:

• Accurate reconstruction of drawing and handwriting from surface electromyography (sEMG) signals is a growing research area.
• Previous methods, such as nonlinear regression, have shown limited success.
• Effective algorithms are essential for reliable myoelectric signal decoding.

Purpose of the Study:

• To propose a novel hybrid algorithm, Kalman Filter-Gene Expression Programming (KF-GEP), for enhanced sEMG-based drawing trace reconstruction.
• To improve the accuracy of transient myoelectric signal decoding.
• To compare the performance of the proposed KF-GEP algorithm against existing Kalman Filter (KF) and Gene Expression Programming (GEP) methods.

Main Methods:

• Developed a hybrid KF-GEP algorithm integrating Gene Expression Programming (GEP) within a Kalman Filter (KF) framework.
• Applied the KF-GEP algorithm to reconstruct fourteen drawn shapes and ten numeric characters.
• Collected sEMG data from five participants during drawing tasks.

Main Results:

• The KF-GEP algorithm demonstrated superior performance in reconstructing drawing and handwriting traces compared to standalone KF and GEP methods.
• The hybrid approach effectively combines the strengths of both KF and GEP for improved decoding accuracy.
• Experimental results validated the algorithm's effectiveness across multiple participants and diverse drawing tasks.

Conclusions:

• The proposed KF-GEP algorithm represents a significant advancement in sEMG-based drawing trace reconstruction.
• This muscle-computer interface technology has potential applications in various practical fields.
• The study contributes valuable insights into advanced myoelectric signal decoding techniques.