Contact Dr Luca Zanotti Fragonara

Areas of expertise

  • Computing, Simulation & Modelling
  • Mechatronics & Advanced Controls
  • Safety, Resilience, Risk & Reliability
  • Sensor Technologies
  • Structures and Materials
  • Systems Engineering

Background

Dr Luca Zanotti Fragonara is Senior Lecturer in Applied Artificial Intelligence in the Centre of Autonomous and Cyber-Physical Systems and in the Aerospace Integration Research Centre. He is director of the Applied Artificial Intelligence MSc course since 2019.

His research is focused on machine learning for structural health monitoring, deep learning and system identification. He is involved in a series of high-impact research projects ranging from autonomous inspections of aerospace structures to structural health monitoring for increasing the resilience of industrial plants.

He matured a wide and unique experience working in monitoring problems applied to civil, mechanical, aerospace and space structures. Moreover, his interests in machine learning allowed him to expand his research activities to other areas, such as perception using camera vision systems, lidars, estimation and sensor fusion.

He also carried out research in the area of non-linear dynamics and finite element modelling.

Current activities

Senior Lecturer, Cranfield University, Centre for Autonomous and Cyber-Physical Systems, Aerospace Integration Research Centre. I am involved in several project linked to:

Autonomous inspections for non-destructive evaluation of aircraft composite materials (http://compinnova.eu/);

Multi User Environment for Autonomous Vehicle Innovation (https://www.cranfield.ac.uk/facilities/mueavi);

Autonomous inspections using drones of composite panels and IRT (MultiAct);

artificial intelligence for perception, object detection, classification;

validation of AI models (https://humandrive.co.uk/);

structural health monitoring for building resilience (https://r.unitn.it/en/dicam/xp-resilience).

I worked on other projects such as:

Vibration-based monitoring of CubeSat;

Land3U (https://www.esa.int/Education/Drop_Your_Thesis/Meet_the_teams_Land3U);

Condition monitoring of foundation piles;

Autonomous inspections of metallic panels using drones.

I teach Sensor Fusion, Deep Learning, AI for Autonomous Systems, Autonomy for Robotic Systems and Statistical Machine Learning.

My wide interests are in deep learning and validation of deep learning models, Structural Health Monitoring of Civil, Mechanical and Aerospace structures, Non-linear dynamics, Finite Element modeling and statistical signal processing.

I am involved in MSc students' supervision during their final thesis, in teaching activities, and research project management activities.

In more detail my major areas of expertise concerns:

Finite Element modelling: linear, non-linear and dynamic models.

Machine Learning algorithms: gradient-descent, logistic regression for multiclass classification, clustering techniques (K-means), Neural Networks, CNNs, R-CNNs, Support Vector Machine (SVM), Anomaly detection algorithms, Principal Component Analysis (PCA).

Numerical methods for Ordinary Differential Equations (ODEs), Volterra series, Associated Linear Equations (ALEs), Differential Algebraic Equations (DAEs).

Nonlinear System identification algorithms: Unscented Kalman Filtering, Particle Filters, Time-Frequency Instantaneous Estimators, Sigma-Point Filters (SPF), Neural Networks, and Autoregressive models (e.g. NARMAX).

Dynamic Identification techniques and experimental modal analysis: time-domain methods (Stochastic Subspace Identification, Eigen-system Realization Algorithm, Neural Networks, Autoregressive models), frequency-domain methods (Frequency Domain Decomposition) and time-frequency domain method (Time-Frequency Instantaneous Estimators, Short-Time Fourier, Choi-Williams, Wigner-Ville based methods, Chirp Fourier Transform, Bi- and Tri-coherence transform).

Nonlinear normal modes and chaos dynamics: shooting methods, arc-length method, pseudo-arc length method.

Global sensitivity analysis techniques (Elementary Effect, Diffused partial-derivatives, Variance-based methods) and meta-modelling techniques.

Clients

Airbus

Airbus DS

European Commission

Transport System Catapult

UK Space Agency

Monitoring and stability evaluation of the Sanctuary of Vicoforte, the largest masonry oval dome in the world, period 2008-2013.

DPC-ReLUIS 2009-2012 (Italian Network of Earthquake Engineering University Laboratories), Research line n. 9, Technologies for Risk Monitoring and Emergency Management. Line 3.1: Development of technologies for monitoring and management of seismic risk. Task 3.1.3: Monitoring.

Program for the seismic and static monitoring and control of the Regina Montis Regalis monumental building, period 2014-2015.

DPC-ReLUIS 2014-2016 (special project 'Seismic Observatory of Structures (OSS) and monitoring').

Publications

Articles In Journals

Conference Papers

Books