EEG Signal Detection Based On Graphene Novel Sensor And Deep Learning Emotional Recognition

Deep learning has promoted the rapid development of intelligent recognition technology,and people's research on the brain is still in the ascendant state.Emotion is a manifestation of brain activity.Emotion has a huge impact on human life and production.How to intelligently recognize human emotions becomes very important.The research direction of this paper is to develop an EEG signal acquisition system capable of stable data acquisition in a non-laboratory environment.The deep learning related algorithm is used to realize the emotion recognition effect based on the graphene electrode data collected in this study.Firstly,this paper develops a new graphene composite electroencephalogram electrode through innovation in microstructure,and attempts to collect EEG signals outside the cerebral cortex through this electrode.Secondly,based on recurrent neural network in EEG signal recognition Widely used in the field,this paper uses two recurrent neural network models to improve,select a model with high recognition accuracy;Finally,use the previously selected network model to compare the dataset collected by graphene electrodes and traditional electrode collection in this study.The recognition rate of the data set output under the same parameters,and then the performance of the graphene electrode in this study.In order to achieve the above objectives,the main work of this paper is as follows:First,this paper firstly found out that the microstructure of the graphene material has excellent electrical properties through a large numbers of hardware performance comparison experiments.It is prepared into a matrix pyramidal graphene electrode;then the flexible circuit board is used as the substrate,and the matrix is used.The pyramidal structure graphene electrode was prepared into an electroencephalogram electrode device by the process developed in this paper.Secondly,the paper firstly encapsulates the EEG electrode device,the independently designed gating circuit module and the purchased signal processing module prepared as a smart hardware wearable system for safe wearing;then the system is used to collect EEG data.Thirdly,in order to verify the data validity of the new graphene composite EEG electrode,this paper uses the deep learning framework Tensorflow based on Anaconda3 as the basis of the algorithm implementation,and uses python3 to optimize the Long and Short Term Memory model and Gated Recurrent Unit model optimization proposed in this paper.Gated Recurrent Unit model is implemented in engineering,and under the verification of the recognition accuracy evaluation index,it is proved that Batch Normalization optimization Gated Recurrent Unit model has higher emotional recognition accuracy.Fourthly,comparing the data collected by the new graphene EEG electrode with the data collected by the traditional electrode in the model determined in the previous chapter,the accuracy of the proposed graphene EEG electrode is valid.