طبقه بندي سيگنال هاي EMG با استفاده از شبكه عصبي موجكي براي رابط انگشت-ربات

Classification of EMG Signals through Wavelet Neural Network for Finger-Robot Interface

The current paper presents Particle Swarm Optimized Wavelet Neural Network (PSOWNN) as a classification method for surface electromyogram (sEMG) pattern classification. According to the literature, a change in the spectrum of surface electromyogram has largely been attributed to the change in muscle conduction velocity. Therefore, such signals are used to command a robot using a WNN classifier. During the experiments, the subjects are instructed by an auditory cue to elicit a contraction from the rest state and hold that finger posture for a period of 5 seconds. For this purpose, two EMG electrodes attached to the human forearm are utilized to collect the EMG data. Time and frequency characteristics such as Number of Zero Crossings (ZC), Autoregressive (AR), and wavelet coefficients are considered as features. And, WNN as a classification method is optimized using particle swarm optimization algorithm. The accuracy of PSOWNN is compared to that of Artificial Neural Network (ANN). The results show an accuracy of 90% for the proposed method, indicating a better performance than ANN in terms of accuracy. Finally, outputs of the best classification method are implemented on a robot.

The current paper presents Particle Swarm Optimized Wavelet Neural Network (PSOWNN) as a classification method for surface electromyogram (sEMG) pattern classification. According to the literature, a change in the spectrum of surface electromyogram has largely been attributed to the change in muscle conduction velocity. Therefore, such signals are used to command a robot using a WNN classifier. During the experiments, the subjects are instructed by an auditory cue to elicit a contraction from the rest state and hold that finger posture for a period of 5 seconds. For this purpose, two EMG electrodes attached to the human forearm are utilized to collect the EMG data. Time and frequency characteristics such as Number of Zero Crossings (ZC), Autoregressive (AR), and wavelet coefficients are considered as features. And, WNN as a classification method is optimized using particle swarm optimization algorithm. The accuracy of PSOWNN is compared to that of Artificial Neural Network (ANN). The results show an accuracy of 90% for the proposed method, indicating a better performance than ANN in terms of accuracy. Finally, outputs of the best classification method are implemented on a robot.