Font Size: a A A

Study On Application Of Portable BCI System

Posted on:2015-06-02Degree:MasterType:Thesis
Country:ChinaCandidate:Y X LiFull Text:PDF
GTID:2298330431978619Subject:Signal and Information Processing
Abstract/Summary:PDF Full Text Request
Brain-computer interface (BCI) is a method of communication or control based on brainwave generated by the brain. It is independent of its normal output pathways, such asperipheral nerves and muscles. With the development of clinical medicine, cognitive science,the technology of computer and communication etc., research on brain-computer interfaceshas become a hot topic in recent years.This paper studies on the subject consists of three major components: the first part is thedesign of electroencephalograph (EEG) acquisition systems, including related hardwarecircuit design and software design; the second part is the EEG signal processing unit,including a processed signal, the signal analysis etc.; the last part is to classify and identify therelevant EEG, and convert the relevant circumstances into a control signal corresponding, thento control the peripheral circuits.The first part in this paper, a new design of portable EEG signal acquisition system isdesigned. Most of the EEG acquisition systems use group of high precision electrical signalamplifier to record and analysis. But the huge input circuit and testing time is difficult toapplication for portable devices. This paper presents an effective chip (ADS1299)implementation of an eight-channel EEG signal acquisition. Compared with the existing system,this design greatly simplifies the front-end circuits and improves the common mode rejectionratio (CMRR). The system has the features of high integration density, good flexibility andpracticability. It is in line with the development trend of modern EEG. Signal acquisitionsystem.The second part introduces the EEG feature extraction algorithm based on wavelet andwavelet packet analysis and the pattern classification method based on support vectormachine (SVM), which effective extract the eigenvector. The eigenvector is characterized bythe coefficient of mean and energy eigenvector. It obtains a better classification effect. Finally,convert the classification information into a control signal; by controlling the small lightseffectively identify the left and the right hand movement. According to the subjects of theexperimental test results show that the system availability is strong and has a high accuracy.
Keywords/Search Tags:Portable, BCI, ADS1299, Feature Selection, Pattern Classification
PDF Full Text Request
Related items