eprintid: 20436
rev_number: 3
eprint_status: archive
userid: 1
dir: disk0/00/02/04/36
datestamp: 2026-01-12 12:18:26
lastmod: 2026-01-12 12:18:26
status_changed: 2026-01-12 12:18:26
type: conference_item
metadata_visibility: show
creators_name: Ooi, Boonyaik Yaik
creators_name: Lim, Jason Jing Wei
creators_name: Liew, Soung Yue
creators_name: Shirmohammadi, Shervin
title: Remote Operation Status Tracking for Manufacturing Machines via Sound Recognition using IoT
ispublished: pub
keywords: Acoustic noise; Manufacture; Sensor nodes; Vibrations (mechanical); Legacy machine monitoring; Machine monitoring; Manufacturing machine; Multilayers perceptrons; Operation status; Operation status tracking; Remote operation; Sound recognition; Status tracking; Tracking accuracy; Internet of things
note: Cited by: 7
abstract: The objective of this work is to track the operation status of legacy manufacturing machines through their vibration and sound emitted during mechanical operations. Although vibration data has been proven successful to track operation status, not all machines allow sensors to be retrofitted for long term operation. Thus, sound-based tracking can potentially be a better alternative for non-intrusive monitoring. Although sound-based tracking can track machines' operation without the need of retrofitting sensors to the machines, the main challenge of using sound-based tracking is that the environment noise will affect tracking accuracy. Therefore, this work attempts to use conventional multilayer perceptron (MLP) models to cut through the noises and track the operation status of machines. In order to have sufficient labelled data for MLP training, we propose an IoT solution that uses temporary battery powered wireless vibration sensor nodes to produce training data with ground truth. The vibration sensor can be discarded once the neural network is trained. Experiments showed that the proposed MLP model can recognize the operation status of two concurrently running machines from a single source of audio and achieves 93.4 tracking accuracy with a standard uncertainty of 0.008 even with untrained dataset with unseen background noises. © 2022 IEEE.
date: 2022
publisher: Institute of Electrical and Electronics Engineers Inc.
official_url: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85134433412&doi=10.1109%2FI2MTC48687.2022.9806481&partnerID=40&md5=79712792f9084bf61571326fb4ca415d
id_number: 10.1109/I2MTC48687.2022.9806481
full_text_status: none
publication: Conference Record - IEEE Instrumentation and Measurement Technology Conference
refereed: TRUE
isbn: 9780879425791; 0879425792; 9781665453837; 9781665483605; 9781467346221; 9781467392204; 0780372182; 1424415411; 0780388798; 078038248X
issn: 10915281
citation:   Ooi, Boonyaik Yaik and Lim, Jason Jing Wei and Liew, Soung Yue and Shirmohammadi, Shervin  (2022) Remote Operation Status Tracking for Manufacturing Machines via Sound Recognition using IoT.  In: UNSPECIFIED.