Research On Group Brain-Computer Interface Technology Based On Motor Imagery

Brain-Computer Interface(BCI)refers to the use of machines by humans,which directly converts the activity information and signal pattern output of the human brain into computer commands without relying on the central nervous system of the human brain or the body movement muscle tissue.By establishing a way of information exchange between the human brain and human external machines and electronic products,a new information exchange method can be carried out between the human brain and the outside world.Motor Imagery(MI)brain-computer interface technology is an active human-computer interaction paradigm.The state of motion has formed a fixed pattern in the brain.It can obtain distinguishable signals without external stimuli,which has become one of many BCI systems.One of the popular research directions.At the same time,there are also problems such as large individual differences,susceptibility to personal status,and low information transmission rate.The Group Brain MI Brain-Computer Interface technology can realize the coordinated control of multiple roles in the execution of BCI,and solve the problem that the EEG signal characteristics generated by the subject during the process of selfmotion imagination in the single person system are not prominent,and the subject is in the test process.Influencing factors such as instability of the body or mental state.In this thesis,combined with the universal brain group algorithm model of Deep Learning,a stable and efficient group brain joint system can be realized.This thesis studies the Brain-Compter Interface technology of the Motor Imagery paradigm,and designs a scheme in which multiple single-person data is merged into group-brain data in the data layer and feature layer,and designs Deep Learning classification models for single-person and group brains.Construction of online realtime control system.It mainly includes the data acquisition part under the group brain paradigm of motor imagery,the serial and parallel data combination parts of the feature layer and the data layer,and the application part of the offline simulation group braincomputer interface motor imagery paradigm in real scenarios.In this thesis,tensorflow2.0-gpu+python3.7 is used to construct the offline classification model of Group Brain-Computer Interface.The online model consists of a group of three male and female subjects with physical and mental health,and a joint test is carried out to verify MI-BCI's online robotic arm sport control.The test results show that the Deep Learning algorithm model designed in this thesis is applied to the Group Brain-Computer Interface multi-person collaborative control system,which can effectively improve the accuracy of real-time transmission control under a large number of tasks,and it is improved by Socket and multi-threading technology.The communication efficiency further guarantees the quality of the final output signal of the sports imagination It is expected that the research results of this paper will bring innovation to the development of Group Brain-Computer Interface technology under the paradigm of motor imagination,and further enrich the theoretical basis of my country's research on this technology,so as to promote the high-quality development of Brain-Computer Interface technology in China.