Comparison of Methods for Determining Activity from Physical Movements

Authors : Mücahit ÇALIŞAN, Muhammed Fatih TALU

Pages : 17-23

Doi:10.2339/politeknik.632070

View : 32 | Download : 11

Publication Date : 2021-03-01

Article Type : Research Paper

Abstract :In this study, the methods which can detect the basic physical movements of a person insert ignore into journalissuearticles values(downward, upward, sitting, stop, walking, running); from inertial sensor insert ignore into journalissuearticles values(IMU); data are evaluated. The performances of classical insert ignore into journalissuearticles values(ANN, SVM, k-NN); and current approaches insert ignore into journalissuearticles values(Convolutional Neural Networks-ESA); to map IMU data to activity classes were compared. A three-stage study was carried out for this aim: 1); data acquisition; 2); creating training/test sets; 3); construction and classification of network architectures. At the stage of data acquisition, to obtain 6 different physical movements from 10 different people, the accelerometer sensor is placed on the persons. Repetitive movements of persons were recorded. At the second stage, the recorded long-term accelerometer data is divided into packages in the form of short-term windows. The training set of classical approaches was constructed by features extracting from each packet data containing one-dimensional acceleration information. The transformation of one-dimensional signals to a two-dimensional image matrix for the training set of the deep learning-based approaches was performed. In the third stage, ANN, SVM, k-NN and CNN architectures were constructed, and classification process was carried out. As a result of the experimental studies, it was found that the accuracy of IMU-activity mapping was 99% with the ANN method and 95% with the CNN method.
Keywords : Physical movement, determining activity, IMU, ANN, CNN

ORIGINAL ARTICLE URL