Citation: Cortes-Aguilar, T.A.;
Cantoral-Ceballos, J.A.;
Tovar-Arriaga, A. Link Quality
Estimation for Wireless ANDON
Towers Based on Deep Learning
Models. Sensors 2022, 22, 6383.
https://doi.org/10.3390/s22176383
Academic Editors: Kelvin K.L. Wong,
Dhanjoo N. Ghista, Andrew W.H. Ip
and Wenjun (Chris) Zhang
Received: 13 July 2022
Accepted: 17 August 2022
Published: 24 August 2022
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2022 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Article
Link Quality Estimation for Wireless ANDON Towers Based on
Deep Learning Models
Teth Azrael Cortes-Aguilar
1,
*, Jose Antonio Cantoral-Ceballos
2,
* and Adriana Tovar-Arriaga
3
1
Centro de Tecnologia Avanzada, CIATEQ A.C., Jalisco 45131, Mexico
2
Tecnologico de Monterrey, School of Engineering and Sciences, Monterrey 64849, Mexico
3
Instituto Tecnologico Jose Mario Molina Pasquel y Henriquez, Jalisco 45019, Mexico
* Correspondence: teth.cortes@zapopan.tecmm.edu.mx (T.A.C.-A.); joseantonio.cantoral@tec.mx (J.A.C.-C.)
Abstract:
Data reliability is of paramount importance for decision-making processes in the industry,
and for this, having quality links for wireless sensor networks plays a vital role. Process and machine
monitoring can be carried out through ANDON towers with wireless transmission and machine
learning algorithms that predict link quality (LQE) to save time, hence reducing expenses by early
failure detection and problem prevention. Indeed, alarm signals used in conjunction with LQE
classification models represent a novel paradigm for ANDON towers, allowing low-cost remote
sensing within industrial environments. In this research, we propose a deep learning model, suitable
for implementation in small workshops with limited computational resources. As part of our work,
we collected a novel dataset from a realistic experimental scenario with actual industrial machinery,
similar to that commonly found in industrial applications. Then, we carried out extensive data
analyses using a variety of machine learning models, each with a methodical search process to adjust
hyper-parameters, achieving results from common features such as payload, distance, power, and bit
error rate not previously reported in the state of the art. We achieved an accuracy of 99.3% on the test
dataset with very little use of computational resources.
Keywords: wireless sensor network; link quality estimation; deep learning; failure detection
1. Introduction
ANDON tower lamps have evolved from simple alarm systems with visual signals in
the traditional factory towards a wireless sensor network (WSN) at the heart of
Industry 4.0
capable of remotely identifying machine status whilst getting information about the work-
shop productivity [
1
,
2
]. The alarm signals for ANDON tower lamps are established by the
international standard IEC 60073:2002 [
3
] in which the red code sends a fault or emergency
stop message, the amber code sends a warning signal for operation in an abnormal process
condition or machine setup, and the green code indicates a normal operation signal, whilst
the blue and white codes represent user-defined signals. These types of systems are readily
available on the market, such as NH-3FV2W from the company Patlite [
4
], TL70 from
Banner Engineering Corp. [5], and SmartMonitor by Werma Signaltechnik [6].
In the industry, the information generated by WSN such as wireless ANDON tower
systems is of paramount relevance because it allows for on-time error correction, applying
containment measures and preventing different problems not possible to anticipate in a
traditional industry. The reliability of such information is even more relevant in Industry
4.0 since it is used for different automation processes aided by the existence of intelligent,
autonomous, and collaborative machines that communicate with each other in real-time
to perform actions based on the environment data [
7
]. In this regard, machine learning
models such as deep neural networks (DNNs) have made remarkable progress that may
have a big impact in the industry with highly reliable analytical solutions [8].
Sensors 2022, 22, 6383. https://doi.org/10.3390/s22176383 https://www.mdpi.com/journal/sensors
