Recognition Of Multi Class Motor Imagery EEG Signals And Its Application In Manipulator Control

Brain-computer interface technology aims to build a new communication channel between the brain and external devices.The brain's intention is directly input to the external device as a control command,thereby helping people with movement disorders to restore communication with the outside world.Widely used in medical rehabilitation,intelligent assistive device control,spelling communication,consumer entertainment and other fields.EEG signal process is indispensable within the analysis and application of brain-computer interface technology,however it still faces challenges like incomplete feature info extraction,low recognition accuracy,and single recognition action class in current applications.During this paper,the associative experimental paradigm is for various motions of the same joint,and therefore the multi-domain features of the record signal within the time domain,frequency domain,and spatial domain is combined,and proposes an EEG signal recognition method based on multi-domain feature fusion.At a similar time,a brain-computer interface system is constructed to realize the control of the external mechanical part,that verifies the sensible significance of the strategy projected during this subject,the most analysis contents ar as follows:(1)Research on feature extraction method of multi-domain feature fusion based on kernel principal component analysis.A feature extraction method that integrates multi-domain features is proposed.In this paper,the AR model method is used to extract the time domain features,the bispectrum analysis method is used to extract the frequency domain features,and the common spatial pattern method is used to extract the spatial domain features of the EEG signals.Then construct a joint feature and generate fusion features by extracting the main components whose cumulative contribution rate is greater than 85%.At the same time,it is verified that the fusion features have good clustering ability.(2)Research on the classification method of optimal twin support vector machine based on bat algorithm.This subject constructs a twin support vector machine optimized based on the bat algorithm.In order to boost the classification performance,the bat algorithmi is employed to optimize the penalty issue of the twin support vector machine.The comparison of different classification method verifies the superiority of the classification methodology planned during this paper.At the same time,through the comparison of datasets,the classification performance of our methodology on public datasets is verified,and also the generalization of the algorithmic program is verified.(3)Design of manipulator control system based on motor imagery EEG signals.The overall structure of the system is built,every part module of the system is meant,the hardware and code of the system ar designed,and therefore the details ar explained.Finally,check the performance of the system and design experiments to test.We analyze the test results and draw conclusions.It is verified that the proposed algorithm has sure sensible worth for the management of the manipulator,and lays a foundation for realizing more actions of the manipulator.