Identification, Diagnosis, Prognosis and Fault Tolreant Control of Complex Dynamic Systems

The increasing complexity of modern control systems requires calls for the development of new identification, diagnosis, prognosis and fault tolerant control methods. The research activities of the group are mainly focused on the following topics. 

Identification of errors-in-variables models and applications. In the above-mentioned complex framework, the identification of dynamic models from data requires taking into account the presence of input, output and process noise. This leads to the so-called errors-in-variables identification. New time-domain and frequency domain algorithms are proposed for both SISO and MIMO systems.

Diagnosis and prognosis of complex mechatronics systems. Among the cutting-edge Industry 4.0 concepts, Diagnostic and Prognostic are gaining importance in industrial automation firms. The growing computational capabilities of computers, both on-board and outside the machines, allows new possibilities in terms of corrective and predictive maintenance. In this context, it is necessary to analyze large quantities of data collected about machine operations and use them to extract useful information for enhancing their production quality and reliability. This leads to the development of new diagnostic and prognostic  procedures based on the combination of system identification tools and machine learning tools. The proposed methodologies are applied on laboratory and industrial case studies. 

Fault diagnosis and fault tolerant control of aerospace systems. The research topic concerns the development of novel fault diagnosis methodologies for Aerial, Ground and Underwater Unmanned Vehicle (UAV, GUV and UUV), general aviation aircraft and spacecraft. Furthermore, by exploiting the information provided by fault diagnosis, such as faults estimates, Fault Tolerant Control (FTC) systems are developed for the different systems above mentioned. Real case application and tests are obtained by means of UAV, GUV, UUV and an ultralight aircraft made available from various research projects.

The research group works in collaboration with the following faculty at DEI: Andrea Tilli (Diagnosis and Prognosis), Andrea Bartolini and Luca Benini (Thermal model identification and fault detection of high performance computing systems), Nicola Mimmo (Fault diagnosis and fault tolerant control of aerospace systems), Lorenzo Marconi (UAV and UGV in Smart Agriculture), Luca De Marchi, Stefano Diciotti, Marco Breschi (Application of system identification to nuclear magnetic resonance (NMR) data).

ERC Fields

  • PE7_1 - Control engineering 
  • PE7_7 - Signal Processing 

 Scientific coordinator: Prof. Roberto Diversi

Faculty

Paolo Castaldi

Associate Professor

Roberto Diversi

Associate Professor

Umberto Soverini

Associate Professor

PhD Students and Research Fellows

Matteo Barbieri

Teaching tutor

Massimiliano Menghini

PhD Student

Teaching tutor