Special Session 1
Predictive Maintenance in Mechatronic Systems: Data-Driven Diagnostics, Condition Monitoring, and Industrial Applications
Organizers
- Jesus Pacheco (main contact) — University of Sonora
- Victor H. Benitez — University of Sonora
- Alberto Fuentes — Ford Motor Co HSAP
- Pratik Satam — University of Arizona
Abstract
Modern industrial environments increasingly rely on predictive maintenance (PdM) to maximize operational availability, minimize unscheduled downtime, and optimize lifecycle management of complex mechatronic assets. By synthesizing advancements in Internet of Things (IoT) infrastructure, edge computing, artificial intelligence, and physical signal processing, PdM transforms traditional reactive and time-based strategies into dynamic, condition-based operational frameworks. This special track focuses on state-of-the-art methodologies and emerging industrial applications across the entire PdM pipeline, from high-fidelity sensor data acquisition and feature extraction (e.g., vibration analysis, thermal profiling, acoustic emissions) to real-time fault detection, isolation, and Remaining Useful Life (RUL) estimation.
This session invites original contributions spanning theoretical foundations, algorithm design, and practical engineering implementations in intelligent condition monitoring and predictive maintenance. We particularly welcome work on unsupervised anomaly detection methods capable of identifying unmodeled degradation and incipient faults in real time, under the non-stationary conditions typical of industrial operation. By bringing together perspectives from control theory, data science, and mechatronic design, the session aims to advance scalable, robust maintenance strategies for next-generation automated systems.
Topics of Interest
We seek original, complete and unpublished work not currently under review by any other journal/magazine/conference. Topics of interest include, but are not limited to:
- Unsupervised ML algorithms for anomaly behavior analysis under non-stationary operating conditions
- Deep learning and physics-informed models for fault prognosis.
- Edge-AI architectures for real-time monitoring.
- Embedded-AI applications for predictive maintenance.
- Industrial IoT integration.
- Anomaly detection under non-stationary operating conditions.
- Quantitative ROI frameworks for large-scale industrial deployment.
- Physics-Informed Neural Networks (PINNs) and hybrid physical-data modeling.
- Remaining Useful Life (RUL) estimation and health degradation prognostics.
Invited Speakers
To be announced.