An EMG hand motion recognition method based on ensemble deep learning

Abstract

The invention discloses an electromyographic signal hand motion recognition method based on integrated deep learning, which firstly performs window segmentation on continuous electromyographic signals. Then, the EMG signal is preprocessed, the frequency domain representation data of EMG signal is extracted by discrete Fourier transform, the time-frequency domain representation data of EMG signal is extracted by discrete wavelet packet transform, and the EMG signal is analyzed by min-max method. Time domain, frequency domain, and time-frequency domain data are normalized. Then, the primary classifier model is designed based on the convolutional neural network, and three primary classifiers are trained using the time domain, frequency domain and time-frequency domain representation data respectively. Finally, a secondary classifier is designed based on the stacking method, and the results generated by the three primary classifications are spliced ​​and used for the training of the secondary classifier. The invention avoids the omission of important information when extracting features manually, and improves the electromyographic signal of the hand. The accuracy of motion recognition.

Description

technical field [0001] The invention relates to an electromyographic signal hand motion recognition method based on integrated deep learning, and belongs to the fields of deep learning, signal processing and pattern recognition. Background technique [0002] EMG is a superposition of bioelectrical signals generated by human muscle movement. Numerous studies have shown that the use of EMG signals can detect human motion intentions. The action recognition of EMG signals can be applied to non-invasive human-computer interaction systems such as hand prosthesis control, wheelchair control, exoskeleton, and virtual interaction. With the development of sensor technology, the collection of EMG signals has become more and more stable. Therefore, the analysis and identification of EMG signals have also received more attention, and improving the accuracy of EMG signal identification has also become an urgent problem to be solved. [0003] At present, most researchers use feature eng...

AI Technical Summary

Problems solved by technology

[0004] Purpose of the invention: In order to overcome the problems of missing important information in the signal and low recognition accuracy in feature extraction under the existing EMG signal recognition technology, the present invention provides a method for hand movement recognition of EMG signals based on integrated deep learning. Using the powerful feature extraction ability of convolutional neural network to extract features of EMG signals

Images