A SVM based classification of EEG for predicting the movement intent of human body

Abstract

In this paper, the EEG (electroencephalograph) signal acquisition equipment is used to collect the EEG signal of human lower limb movement intention. This paper firstly analyzes α waveform and β waveform, which can most reveal the intentions of human body movement. Then, wavelet transform is used for noise removal, filter and feature extraction. This paper also has described the theory of Support Vector Machine (SVM), and one-to-one SVM method is used for the classification of EEG of six different movement patterns. Finally through the experimental verification, the validity of the proposed research method is demonstrated. The experiment has shown a better judging result, in which the average recognition rate is 78.9%.