MR V V RAMALINGAM,DR S MOHAN,DR A PANDIAN,DR V SUGMARAN,MRS B REBECCA JEYAVADHANAM
申请号：
IN201741020227
公开号：
IN201741020227A
申请日:
2017.06.09
申请国别(地区):
IN
年份:
2017
代理人:
摘要：
Whenever there is a hindrance to day to day activities of human being due to amputation, there arises a need for rehabilitation and its best served by the application of machine learning. According to the desires of the human beings, the function towards the activities of the artificial limb movements which are expected to perform. This invention is about controlling an artificial limb which will be capable of performing movements such as Finger open (Fopen), Finger close (Fclose), Wrist counter clockwise (WCCW) and Wrist clockwise . (WCW) rotation. According to the desires of the human beings, the function towards the activities of the artificial hand movements which are expected to perform. The source of control for this artificial limb will be the EEG signal recorded directly from the human brain activity created through thought process. According to the desires of the human beings, the function towards the activities of the artificial hand movements which are expected to perform. This was conducted by recording EEG signals from 27 different subjects with sound health. Focus has been made on three important phases of machine learning which is feature extraction, feature selection and feature classification. This work extracts Discrete Wavelet features from EEG signals and uses them as input for classification. Feature selection process was done with the help of C4.5 Decision Tree algorithm. Classification of the extracted features was carried out using Logistic Model Tree (LMT) algorithm. This proposed system identifies the best wavelet features that can be utilized to control the artificial limb movements system.
