On July 10, 2025, Teoresi MedTech will present a scientific paper at the international conference IEEE SAS – Sensors Applications Symposium, showcasing the SPS (Smart Predictive System) project, an innovative solution for predictive maintenance of electric motors in industrial environments. The system is based on self-sustainable IoT sensor nodes and an unsupervised learning algorithm for automatic detection of operating conditions and early identification of anomalies. The paper will be presented by Antonio Cancilla (Teoresi MedTech), co-authored with Guido Comai (Teoresi MedTech) and Tommaso Polonelli (ETH Zürich).
In the transition toward intelligent and sustainable production models, Teoresi MedTech stands out for its advanced solutions in predictive maintenance of electric motors, specifically designed for the manufacturing and process automation industries. The SPS project integrates wireless self-powered sensor technology, machine learning algorithms, and scalable software architectures, with the aim of reducing operational costs and unplanned downtime.
Antonio Cancilla’s contribution to IEEE SAS 2025, taking place in Newcastle upon Tyne (UK), is part of the conference’s scientific program and will explore a system based on:
- Wireless self-sustainable sensors, designed for retrofit installation on existing motors, powered through thermoelectric energy harvesting;
- An Incremental DBSCAN clustering algorithm, capable of continuously and adaptively learning from the motor’s operating conditions—without requiring any prior training;
- A software infrastructure for data acquisition, normalization, dimensionality reduction (PCA), and clustering of multisensor data (vibration, acoustic signals, and magnetic field).
The system has been validated on three-phase electric motors ranging from 11 kW to 160 kW in real industrial environments, achieving 97% accuracy in anomaly classification and proving its ability to adapt to previously unknown operating states.
Paper Details
Title: Unsupervised Predictive Maintenance on Industrial Electric Motors Based on Self-Sustainable IoT Wireless Sensor Nodes
Presenter: Antonio Cancilla, Teoresi MedTech
Co-authors: Guido Comai (Teoresi MedTech), Tommaso Polonelli (ETH Zürich)
Conference: IEEE Sensors Applications Symposium