Leg amputees motion pattern recognition based on principal component analysis and BP network

Author

Liu Lei ; Yang Peng ; Liu Zuojun ; Geng Yanli ; Zhang Jun

Author_Institution

Sch. of Control Sci. & Eng., Hebei Univ. of Technol., Tianjin, China

fYear

2013

fDate

25-27 May 2013

Firstpage

3802

Lastpage

3804

Abstract

The problem is the poor gait recognition accuracy for existing power-type prosthetic knee control. In order to improve the accuracy of the classification of prosthetic control system, the paper first analysel the signal acquisition of MMA7361L acceleration sensor and ENC-03 Gyro using the principal component analysis (PCA). It is used for the feature extraction, finally the BP neural network is used for training and testing. The experiment results show that this method can recognize lower limb prosthesis walking uphill, downhill, up and down stairs and different movement pattern recognition quickly and effectively.

Keywords

acceleration measurement; artificial limbs; backpropagation; feature extraction; gait analysis; gyroscopes; medical signal detection; neural nets; pattern classification; principal component analysis; BP neural network; ENC-03 Gyro; MMA7361L acceleration sensor; PCA; feature extraction; gait recognition accuracy; leg amputees motion pattern recognition; lower limb prosthesis down stair walking recognition; lower limb prosthesis downhill walking recognition; lower limb prosthesis up stair walking recognition; lower limb prosthesis uphill walking recognition; movement pattern recognition; power-type prosthetic knee control; principal component analysis; prosthetic control system classification accuracy improvement; signal acquisition; Acceleration; Biological neural networks; Feature extraction; Pattern recognition; Principal component analysis; Prosthetics; Accelerometer; BP algorithm; Gyroscope; Principal Component Analysis;

fLanguage

English

Publisher

ieee

Conference_Titel

Control and Decision Conference (CCDC), 2013 25th Chinese

Conference_Location

Guiyang

Print_ISBN

978-1-4673-5533-9

Type

conf

DOI

10.1109/CCDC.2013.6561611

Filename

6561611