Home // ACCSE 2018, The Third International Conference on Advances in Computation, Communications and Services // View article

Machine Learning-Based Object Detection System Using PIR Sensor

Authors:
Dong Hyun Kim
Jung Bin Park
Jong Deok Kim

Keywords: machine learning; object detection system; PIR sensor

Abstract:
An intrusion prevention system using a digital Pyroelectric Infra-Red (PIR) sensor produces an error with an object, not a human. To solve this error, this research suggests an analog PIR sensor and an object detection system using machine learning. The analog PIR sensor provides an output based on various voltage scales within a certain area rather than producing binary outputs using a threshold value. From samples of an analog signal attained by using an analog PIR sensor, a Fast Fourier Transform (FFT) processed frequency is produced and used as a feature vector of the Artificial Convolutional Neural Network (CNN). The artificial CNN then studies the signal patterns of human motion and animal motion and detects whether it is a human or animal that intruded.

Pages: 11 to 14

Copyright: Copyright (c) IARIA, 2018

Publication date: July 22, 2018

Published in: conference

ISSN: 2519-8459

ISBN: 978-1-61208-658-3

Location: Barcelona, Spain

Dates: from July 22, 2018 to July 26, 2018