Real-time EMG Pattern Recognition for Hand Movements
Author Information
Author(s): Khezri Mahdi, Jahed Mehran
Primary Institution: Sharif University of Technology, Electrical Engineering Department, Biomedical Engineering Group, Tehran, Iran
Hypothesis
Can a real-time intelligent pattern recognition algorithm improve the accuracy of surface EMG signal classification for hand prosthesis control?
Conclusion
The study demonstrates that the ANFIS real-time learning method can achieve a 96.7% average accuracy in recognizing hand movements, which is superior to previous methods.
Supporting Evidence
- The ANFIS approach achieved a 96.7% average accuracy in recognizing hand movements.
- The study utilized a hybrid method combining back-propagation and least mean square for training.
- Real-time learning was incorporated to adapt to user feedback during the classification process.
- The system was tested on six distinct hand movements with a total of 600 samples.
Takeaway
This study shows that a smart system can help people with prosthetic hands move them better by understanding their muscle signals.
Methodology
The study used an adaptive neuro-fuzzy inference system (ANFIS) with real-time learning to classify EMG signals from six hand movements.
Potential Biases
Potential bias in user evaluation and training data collection.
Limitations
The study may not account for variations in EMG signals across different individuals or conditions.
Participant Demographics
Four healthy subjects participated in the study.
Statistical Information
P-Value
0.0001
Confidence Interval
Not specified
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website