Research And Implementation Of Emotion Recognition Technology Based On EEG

With the development of brain-computer interface technology,emotion recognition based on EEG has been favored and valued by researchers.Research on EEG emotion recognition has injected new vitality into the development of artificial intelligence.This thesis focused on the two aspects of EEG signal removal and pattern recognition in EEG signals,and designed a service robot system based on EEG emotion recognition,which realized the integrated application of the algorithm in the system.It has a certain theoretical significance and practical application value.Firstly,this thesis summarized the EEG signals,emotion models and the current research status at home and abroad,and the system of emotion recognition based on EEG signals is designed.Then the EEG signals acquisition system is introduced and the EEG data is collected.Finally,the common EEG signals analysis methods are analyzed,and it is determined that the deep learning method is applied to the EEG signal processing.Secondly,to solve the problem that the traditional methods of removing the electroencephalogram artifacts requires electrooculogram signals as reference signals,and it is easy to cause the loss of important information of EEG signals or the incomplete removal of artifacts.The joint of sparse autoencoder and denoising autoencoder are explored to solve the above problem in this paper,and then an electroencephalogram artifact removal algorithm based on stacked sparse denoising autoencoder is proposed.The method is divided into an offline stage and an online stage,where the offline stage completes the training of the model,and the online stage removes the electroencephalogram artifacts in the EEG signals.Experimental results show that the proposed method can effectively remove electroencephalogram artifacts from EEG signals,and is better than the other methods.Thirdly,to solve the problem of low emotional recognition rate of EEG signals,this thesis comprehensively considers the rich time-frequency information of EEG signals and the spatial information between channels,and proposes a three dimension convolutional bidirectional recurrent neural network(3DCNN)pattern recognition method.The method takes the time-frequency-channel three-dimensional data as input,which introduces three dimension convolutional neural network and bidirectional recurrent neural network to extract features and classify.Experimental results show that this method has higher average recognition rate than other methods.Finally,a service robot system based on EEG emotion recognition is designed to complete the integrated application of the algorithm.The system controls the service robot by collecting the subject's EEG signals,and the service robot can respond accordingly according to different emotions,thereby achieving more friendly human-computer interaction.The experimental comparison with other algorithms shows the effectiveness of the proposed algorithm in practical applications.