The Research Of Lower Limb Rehabilitation System Based On Eeg In Stroke

Modern medical rehabilitation theory proves that the training can effectively restore motor control to stroke survivors and it will have better effects when stroke survivors participate in. The study of neurorehabilitation in stroke using EEG-based brain-computer interface (MI-BCI) with robotic rehabilitation provides evidence that:a mild stroke patient can effectively controlled robotic rehabilitation based MI-BCI and it is effective in restoring limbs motor function in stroke .In this paper, a Limb recovery system based MI is developed for regular training of neurological rehabilitation for stroke patients.We study the pattern recognition of motor imagery EEG of the left hand, right hand in this paper.The pattern recognition of EEG includes feature extraction and classification. First of all, power spectral density analysis and bispectrum analysis were used to motor imagery EEG, according to the results of analysis, power spectral density values nearby 10Hz, 20Hz and the maximum of bicoherent coefficient were extracted as features, then we use the support vector machines (SVM) as classifier, the accuracy was 87%.Rehabilitation system hardware structure including: micro-controller section, electromagnetic valve steering, and exchange pump, filling put gas device and pressure sensor parts. This paper adopts ARM Cortex-M3 micro controller, which combines received EEG control signal and pressure sensor feedback signal, to control the pump and electromagnetic valvel, and then realize rehabilitation training for stroke patients. This paper selects embedded apmsμC/OS-Ⅱoperating system as the recovery system software development platform, it combines the software to hardware and controls rehabilitation device according to established target motion by writing program.