Research On Brain-computer Interface System Based On FNIRS

Brain-computer interface based on functional near infrared spectroscopy(fNIRS)is a new human-computer interaction method.The brain computer interface based on fNIRS is an important brain computer interface,which has a wide application prospect.The existing brain computer interface based on fNIRS is mainly used to recognize the motion imagination of different limbs,and the generation of external device control instructions is limited.In order to make this kind of brain computer interface practical,online task recognition is needed.In addition,the regulation of neural feedback on the brain mechanism of motor imagination has been neglected in the previous studies,so as to improve the control accuracy of motor imagination.Based on the existing research,this thesis mainly does the following research:(1)Research on recognition of three imaginative movements of a single limb based on fNIRS.In this study,the three movements of passing,stopping and shooting in football were used as experimental sports imagination tasks.During the task,the subjects imagined three movements of the right foot according to the prompts.In the research,the correlation coefficient between different task signals is extracted as a new feature of recognition.Finally,support vector machine is used to classify the three tasks,and the classification accuracy is 78.89% ± 6.161%.In addition,the common signal mean and signal slope are extracted for task recognition.The results show that fNIRS with different motion imagination of a single limb has separability,which is expected to add new control commands to fNIRS-BCI.In addition,the correlation coefficient can be used as an effective feature.(2)An on-line brain computer interface system based on fNIRS is built.In this part,the on-line brain computer interface system is built by using the existing functional near-infrared light equipment.Four parts of the system are designed: acquisition of real-time task data,real-time signal processing and feature extraction,real-time classification and recognition,and real-time instruction sending.After testing,the on-line system can be used for the on-line research of brain computer interface of functional near infrared spectroscopy.(3)The research of brain computer interface based on fNIRS.In this part,online fNIRS-BCI research is carried out using the established experimental platform.The experiment designed an experimental paradigm about right hand clenching and right forearm moving to the left,and tested the performance of the online system built in this thesis.The accuracy ofclassification was 72.25% ± 0.004%.Results the good performance of the developed on-line brain computer interface system based on fNIRS was verified,and the feasibility of on-line brain computer interface was verified,which is expected to be applied in practice.(4)Based on fNIRS neural feedback control movement imagination action research.In the experiment,the left and right hand motion imagination and the two feet motion imagination respectively correspond to the control commands of robot turning left 45 °,right 45 ° and forward one step.During the experiment,the subjects were randomly divided into experimental group and control group.By comparing the classification accuracy before and after neural feedback training with that of the experimental group and the control group.The experimental results show that neural feedback training can improve the control and shaping of motor imagery and regulate the mechanism of brain motor imagery.