Home People Gian A. Susto Research

Machine Learning for Industry 4.0

The so-called '4th Industrial Revolution' is making distruptive changes in manufacturing, product development and business process management. This revolution is centered around the availability of data and their exploitation. Machine Learning provides the methodological tools to translate data into information and value.

I've been working in several data-centered application technologies:

 

Predictive Maintenance

Anomaly/Fault Detection 

Run-to-Run Control

Soft Sensors/Virtual Metrology

Dynamic Sampling
Review Publications:

 

 

Human Activity & Gesture Recognition

Recognizing human behaviour in smart environment or with wearable devices is a challenging task with many applications areas: home automation, health care, gaming and sports.

The challanges in this area are related to dealing with streams of complex data and low resources devices.

- A multivariate symbolic approach to activity recognition for wearable applications
- A Symbolic Approach to Human Activity Recognition

 

 

 

Control of Partial Differential Equations

Partial Differential Equations (PDEs) are used to describe natural phenomena like sound, heat, fluid flow, electrodynamics, elasticity, etc. How to control system described by PDEs is still an opening challenge in the control system reseach area.

In the following work it has been studied how to control a system in which the actuator or the sensor dynamics is described by a PDE: