
Giuseppe NUNNARI
Giuseppe Nunnari is a former Full Professor of Automatic Control (SSD IINF-04/A (Automatica)) at the Department of Electrical, Electronic, and Information Engineering (DIEEI) of the University of Catania, Italy, where he has taught and conducted research since 1992. Retired since November 1, 2025, due to age limits, he currently continues his academic and educational commitment as an Adjunct (Contract) Professor of Industrial Automation within the same department. He received his Laurea degree (M.S. equivalent) in Electrical Engineering with honors from the University of Catania in 1979.
He began his research career focusing on digital signal filtering and the modeling of human motor systems (collaborating with the CNR and Politecnico di Milano). In 1983, he joined the International Institute of Volcanology of the Italian National Research Council (CNR) as a tenured researcher, where he pioneered the application of systems theory and artificial intelligence to geophysical and environmental modeling. In 1992, he joined the University of Catania as an Associate Professor, becoming extraordinary professor in 2001 and attaining the rank of Full Professor in 2004.
An active and prolific researcher, Professor Nunnari is the author or co-author of 4 scientific books (including monographs published by Springer-Verlag) and several book chapters. His research portfolio comprises over 80 articles published in international peer-reviewed indexed journals and more than 130 papers in international conference proceedings. His scientific contributions span several key domains of cybernetics and control engineering:
1. Intelligent Systems & Soft Computing: Developing neural network and fuzzy logic models for the identification and advanced controller design of highly non-linear industrial processes.
2. Environmental & Geophysical Cybernetics: Applying machine learning, neural networks, and multivariate data analysis to geological data, lava fountain simulations at Mt. Etna, and volcanic warning systems (such as the THEODOROS system). He was notably among the absolute pioneers in deploying Global Positioning System (GPS) techniques for the scientific monitoring of volcanic ground deformations at Mt. Etna since the early 1990s.
3. System Realization & Reduction: Devising analytical methodologies for stable, low-order balancing and approximations of high-order Linear Time-Invariant (LTI) systems.
4. Computer Vision: Designing Cellular Neural Network (CNN) simulators for real-time video and image processing in industrial and natural monitoring.
He has been a key scientific partner in major European framework programs (including TECVOLC, APPETISE, ROBOVOLC, and DICTAM) and has established research collaborations with leading international institutions such as the University of New Mexico, Institut de Physique du Globe de Paris, and the Czech Academy of Sciences. Beyond his university courses, he actively serves as a specialized instructor for premium industrial master programs and regional technical academies, preparing the next generation of mechatronics, automation, and Industry 4.0 professionals.
The most recent research activity (2015-Present) dealt with:
- Study of new properties of Multi-Input Multi-Output Linear time-invarian systems. In particular, I have studied the relationship between the Hanke singular values and characteristic values of the cascade system between a MIMO and a inner system in the continuous time domain.
- Characterization of a volcanic system as a Self-Organized Critical (SOC). In particular, analyzing a large database with most of all the known volcanic eruptions, I have contributed to determie that the duration of eruptions can be described by a universal distribution which characterizes eruption duration dynamics.
- Assessment of the ongoing activity of volcanoes is order to reduce volcani risks. To this purpose, I have applied various kinds of Machine Learning Approaches to various kind of data set recorded in the Mount Etna Area.
- Study of solar radiation prediction models using deep learning techniques,
- Study of monthly sales prediction models based on recorded time series.