EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A Multimodal TinyML-Based Predictive Maintenance Architecture for Industrial IoT in the 6G Era

Carlos Exequiel Garay iD, Fernando Alberto Miranda Bonomi iD, Gonzalo Nicolás Mansilla, Mariano Fagre iD, Sergio Gustavo Guzmán, Pablo Alberto Ritorto, Franco Ismael Perez, Marcos Katz

Sensors · 17.07.2026 · 26 (14) · 4536

DOI: 10.3390/s26144536 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 17.07.2026. Verfasst von Carlos Exequiel Garay und 7 weiteren Autor:innen. Beteiligte Einrichtung: CIASUR (Centro de Investigación de Atmósfera Superior y Radiopropagación), Facultad Regional Tucumán (FRT), Universidad Tecnológica Nacional (UTN), R…. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 80 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: CIASUR.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Predictive maintenance (PdM) is central to Industry 5.0 strategies for reducing unplanned downtime in rotating machinery. This work proposes and evaluates, as a proof of concept on a controlled single-machine testbed, a multimodal TinyML edge architecture for PdM designed to remain compatible across the application plane’s evolution toward sixth-generation (6G) networks. Three complementary modalities run local inference on commercial off-the-shelf smart sensor nodes—vibration, acoustic, and thermography—with an embedded gateway bridging per-modality decisions to a serverless cloud back-end. Using real vibration data from a controlled static-unbalance protocol, five anomaly-detection model variants, operating on ten frequency-independent time-domain features extracted from 6 s windows, are benchmarked on the actual Cortex-M4F target; the INT8-quantized fully connected autoencoder, scored by per-window reconstruction error, reaches F1 = 0.9807 with 254 µs inference latency and a 6056 B Flash footprint, well within the microcontroller budget. In a second acquisition session with the remounted sensor, the frozen model retains perfect fault recall, and a short per-installation healthy-baseline recalibration restores F1 = 0.975 without any weight retraining. The acoustic modality is classified in-sensor on log-Mel filterbank energies by the Syntiant NDP120 neural coprocessor, and the thermographic modality by a lightweight binary CNN on 96 × 96 px frames. A preliminary intra-session late-fusion analysis suggests that a logistic-regression meta-learner over the three modality confidence scores can improve on single-modality baselines when no single modality already saturates, motivating multimodal sensing primarily for robustness and redundancy. An end-to-end latency experiment shows that the cloud-uplink leg dominates the budget (79–88%), establishing edge-first inference as a necessary condition for 6G URLLC gains to be observable at the application level. All experiments are conducted over Wi-Fi and MQTT with no 5G or 6G radio, so 6G compatibility is presented as a forward-looking roadmap rather than a tested capability.

Schlagwörter

Stichwörter der Autor:innen

6GIndustry 5.0TinyMLURLLCanomaly detectionexperimental validationmultimodal sensingpredictive maintenancesensor fusionthermography

Autor:innen und Institutionen

  • CIASUR (Centro de Investigación de Atmósfera Superior y Radiopropagación), Facultad Regional Tucumán (FRT), Universidad Tecnológica Nacional (UTN), Rivadavia 1050, San Miguel de Tucumán 4000, Argentina.
  • Laboratorio de Telecomunicaciones (LTC), Departamento de Electricidad, Electrónica y Computación, Facultad de Ciencias Exactas y Tecnología (FACET), Universidad Nacional de Tucumán (UNT), San Miguel de Tucumán 4000, Arg…
  • Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), San Miguel de Tucumán 4000, Argentina.
  • Laboratorio Área IV-Termología, Departamento de Mecánica, Facultad Regional Tucumán (FRT), Universidad Tecnológica Nacional (UTN), Rivadavia 1050, San Miguel de Tucumán 4000, Argentina.
  • Centre for Wireless Communications, University of Oulu, 90014 Oulu, Finland.

ORCID-Profile: Carlos Exequiel Garay, Fernando Alberto Miranda Bonomi, Mariano Fagre

Förderung

  • CIASUR

Zitieren

APA

Garay, C. E., Miranda Bonomi, F. A., Mansilla, G. N., Fagre, M., Guzmán, S. G., Ritorto, P. A., Perez, F. I., & Katz, M. (2026). A Multimodal TinyML-Based Predictive Maintenance Architecture for Industrial IoT in the 6G Era. Sensors, 26(14), 4536. https://doi.org/10.3390/s26144536

Vancouver

Garay CE, Miranda Bonomi FA, Mansilla GN, Fagre M, Guzmán SG, Ritorto PA, et al. A Multimodal TinyML-Based Predictive Maintenance Architecture for Industrial IoT in the 6G Era. Sensors. 2026;26(14):4536. doi:10.3390/s26144536

BibTeX
@article{Garay2026,
  author = {Garay, Carlos Exequiel and Miranda Bonomi, Fernando Alberto and Mansilla, Gonzalo Nicolás and Fagre, Mariano and Guzmán, Sergio Gustavo and Ritorto, Pablo Alberto and Perez, Franco Ismael and Katz, Marcos},
  title = {A Multimodal TinyML-Based Predictive Maintenance Architecture for Industrial IoT in the 6G Era},
  journal = {Sensors},
  year = {2026},
  volume = {26},
  number = {14},
  pages = {4536},
  doi = {10.3390/s26144536}
}
Kostenlos starten →