Application Research Of Online Brain-controlled Robot System Based On BCI

Brain-computer interface(BCI)is a revolutionary human-computer interaction,of which the one based on Motor Imagery(MI)is a very important brain-computer interaction.The brain-computer interface technology involves multidisciplinary cross-cutting research.It “make use of the brain”,namely bypass the peripheral nerves and muscles of the brain based on“understanding the brain” to realize a direct communication and control with external devices through the brain signals.As a leading international research hotspot,BCI is not only strategically expected to be used for defense and military purposes but also to provide the disabled with a new channel,a direct access to external devices or robots so as to improve their quality of life.It is more likely to provide normal people with new ways of controlling external devices or robots under special conditions to increase the quality of their lives.This paper designs and builds a robot brain control system based on BCI as well as provides a platform for training subjects to improve their sports imagination.It first provides a wireless brainwave collection earphone for the motor imagery brain region,which has long service life and stable impedance and is not limited by the subject's range of activity.It is expected to be applied in practice.Secondly,this paper adopts a detailed research on the on-line brain-computer interface system based on MI and analyzes the experiment result of pretreatment methods,feature extraction and pattern recognition methods,which proves the effectiveness of the pretreatment method used in this paper and provides a good approach for the following experiment.Then it conducts a classified study of four categories of MI based on the algorithm of Common Spatial Pattern(CSP)and Ridge Regression(RR)of the narrow EEG signals and makes a comparison of the accuracy between two feature extraction algorithms and three classification algorithms.Through the comparison and combination,the algorithm of optimal feature extraction and classification is obtained,paving a way for the following related research and laying a foundation for the online four classification research.Finally,this paper designs a three-dimensional robot motor system based on multi-task motor imagery EEG and introduces the components of the system in detail.It also describes related devices to meet the demands of online control.By using of the virtual robot BCI system,the experiment of online motor imagery EEG is completed,which proves the effectiveness of the system.At the same time,the system also provides a good platform for training subjects to improve their motor imagery.