Giancarlo Ferrari Trecate

Giancarlo Ferrari Trecate
Prof.
Giancarlo Ferrari Trecate
PI
Executive Committee Member
Understanding the potential and the limits of pervasive control algorithms is crucial for building a more sustainable and safer future.

Giancarlo Ferrari-Trecate received the Ph.D. degree in Electronic and Computer Engineering from the Universita' degli  Studi di Pavia in 1999. Since September 2016 he is Professor at EPFL, Lausanne, Switzerland. In spring 1998, he was a Visiting Researcher at the Neural Computing Research Group, University of Birmingham, UK. In fall 1998, he joined as a Postdoctoral Fellow the Automatic Control Laboratory, ETH, Zurich, Switzerland. He was appointed Oberassistent at ETH, in 2000. In 2002, he joined INRIA, Rocquencourt, France, as a Research Fellow. From March to October 2005, he was researcher at the Politecnico di Milano, Italy. From 2005 to August 2016, he was Associate Professor at the Dipartimento di Ingegneria Industriale e dell'Informazione of the Universita' degli Studi di Pavia.

His research interests include scalable control, microgrids, networked control systems, hybrid systems and machine learning.

Giancarlo Ferrari Trecate was the recipient of the Researcher Mobility Grant from the Italian Ministry of Education, University and Research in 2005. He is currently a member of the IFAC Technical Committees on Control Design and Optimal Control, and the Technical Committee on Systems Biology of the IEEE SMC society. He has been serving on the editorial board of Automatica for nine years and of Nonlinear Analysis: Hybrid Systems.

Scientific Publications

Published
Minimal Regret State Estimation of Time-Varying Systems
Proceedings of IFAC World congress 2023
Published
Follow the Clairvoyant: An Imitation Learning Approach to Optimal Control
Proceedings of IFAC World congress 2023
Published
Stable Linear Subspace Identification: A Machine Learning Approach
ArXiv
Published
Data-Enabled Predictive Control for Empty Vehicle Rebalancing
European Control Conference (ECC 23)
Published
Universal Approximation Property of Hamiltonian Deep Neural Networks
IEEE Control System Letters
Vol 7 Pages 2689-2694
Published
Regularization for distributionally robust state estimation and prediction
IEEE Control System Letters
Vol 7 Pages 2713 - 2718
Published
Safe Zeroth-Order Convex Optimization Using Quadratic Local Approximations
Proceedings of ECC2023
Published
Maximum likelihood estimation of distribution grid topology and parameters from smart meter data
Grid edge technologies 2023
Published
Unconstrained Parametrization of Dissipative and Contracting Neural Ordinary Differential Equations
62nd IEEE Conference on Decision and Control
Published
Hamiltonian Deep Neural Networks Guaranteeing Nonvanishing Gradients by Design
IEEE Transactions on Automatic Control
Vol 68 No 5 Pages 3155-3162
Published
Optimal control configuration in distribution network via an exact OPF relaxation method
2022 IEEE 61st Conference on Decision and Control (CDC)
Pages 5698-5704
Published
Robust online joint state/input/parameter estimation of linear systems
2022 IEEE 61st Conference on Decision and Control (CDC)
Published
Physically Consistent Neural ODEs for Learning Multi-Physics Systems
Proceedings of IFAC World congress 2023
Published
Bayesian Error-in-Variables Models for the Identification of Distribution Grids
IEEE Transactions on Smart Grid
Vol 14 No 2
Published
Idle-vehicle rebalancing coverage control for ride-sourcing systems
2022 European Control Conference (ECC)
Pages 1970-1975
Published
Finite-sample-based Spectral Radius Estimation and Stabilizability Test for Networked Control Systems
2022 European Control Conference (ECC)
Pages 2087-2092
Published
Robust Classification Using Contractive Hamiltonian Neural ODEs
IEEE Control Systems Letters
Vol 7 Pages 145-150
Published
Optimal Adaptive Droop Design via a Modified Relaxation of the OPF
IEEE Transactions on Control Systems Technology
Vol 31 No 2 Pages 497-510
Published
Distributed Neural Network Control with Dependability Guarantees: a Compositional Port-Hamiltonian Approach
Proceedings of The 4th Annual Learning for Dynamics and Control Conference
Vol 168 Pages 571-583
Published
Near-Optimal Design of Safe Output Feedback Controllers from Noisy Data
IEEE Transactions on Automatic Control
Pages 1-16
to Appear
Optimal droop control placement in distribution network via an exact OPF relaxation method
61st IEEE Conference on Decision and Control (CDC 2022)
to Appear
Neural System Level Synthesis: Learning Over All Stabilizing Policies for Nonlinear Systems
61st IEEE Conference on Decision and Control (CDC 2022)
Published
Safe Control with Minimal Regret
Proceedings of The 4th Annual Learning for Dynamics and Control Conference
Vol 168 Pages 726-738
Published
Bayesian Methods for the Identification of Distribution Networks
2021 60th IEEE Conference on Decision and Control (CDC)
Pages 3646-3651
Published
Neural Energy Casimir Control for Port-Hamiltonian Systems
arXiv
A Unified Framework for Hamiltonian Deep Neural Networks
Proceedings of the 3rd Conference on Learning for Dynamics and Control
Vol 144 Pages 275 – 286
Hamiltonian Deep Neural Networks Guaranteeing Non-vanishing Gradients by Design
IEEE Transactions on Automatic Control
A Behavioral Input-Output Parametrization of Control Policies with Suboptimality Guarantees
60th IEEE Conference on Decision and Control (CDC21)

Research projects

Title
Principal Investigators

Layout optimization for decentralized operation of a network of power converters

Summary

The project addresses the question of optimal system layout for facilitating decentralized control and optimal operation of a network of power converters with generators, storages and loads. Special emphasis is placed on the system capability of operating in a decentralized fashion to meet specific local control objectives (e.g. keeping the local voltage within prescribed limits) and system-wide control objectives (e.g. fair proportional load sharing among the inverters, provision of ancillary services). Sensitivity analysis will enable to define the boundaries of a completely decentralized structure and where adding a hierarchical level enables a significant performance improvement.

Layout optimization for decentralized operation of a network of power converters

The project addresses the question of optimal system layout for facilitating decentralized control and optimal operation of a network of power converters with generators, storages and loads. Special emphasis is placed on the system capability of operating in a decentralized fashion to meet specific local control objectives (e.g. keeping the local voltage within prescribed limits) and system-wide control objectives (e.g. fair proportional load sharing among the inverters, provision of ancillary services). Sensitivity analysis will enable to define the boundaries of a completely decentralized structure and where adding a hierarchical level enables a significant performance improvement.

149
608284ed-f17e-4590-9f04-69f85324825e

Structured data-based distributed control

Summary

Cyberphysical systems are composed by subsystems linked through physical interconnections and communication channels enabling distributed control architectures. By leveraging recent results on optimal control, we will develop distributed algorithms for the data-driven generation of networked regulators adapted to the communication topology. Special attention will be devoted to the decentralization of the design process, so as to facilitate the addition and removal of subsystems in a plug-and-play fashion. As benchmark examples, we will consider energy systems and manufacturing processes, which are key applications in the NCCR.

Structured data-based distributed control

Cyberphysical systems are composed by subsystems linked through physical interconnections and communication channels enabling distributed control architectures. By leveraging recent results on optimal control, we will develop distributed algorithms for the data-driven generation of networked regulators adapted to the communication topology. Special attention will be devoted to the decentralization of the design process, so as to facilitate the addition and removal of subsystems in a plug-and-play fashion. As benchmark examples, we will consider energy systems and manufacturing processes, which are key applications in the NCCR.

121
3a8de74d-4208-4166-8063-9d103f7edaf8

ODE networks for data-based control

Summary

In spite of several success stories, standard deep networks  can lead to unstable forward propagation or suffer from an ill-posed learning process, which might result in vulnerability to adversarial attacks. In this project we will study deep network architectures based on ODEs for mitigating these problems. Furthermore, we will investigate the use of these networks for parametrizing regulators that can be tuned from available data. The new methods will be tested for the control of inverters, smart loads and microgrid clusters, which are relevant for the energy systems considered in the NCCR.

ODE networks for data-based control

In spite of several success stories, standard deep networks  can lead to unstable forward propagation or suffer from an ill-posed learning process, which might result in vulnerability to adversarial attacks. In this project we will study deep network architectures based on ODEs for mitigating these problems. Furthermore, we will investigate the use of these networks for parametrizing regulators that can be tuned from available data. The new methods will be tested for the control of inverters, smart loads and microgrid clusters, which are relevant for the energy systems considered in the NCCR.

120
bba0c33c-021f-4dbb-848c-3b2f57e14a0b

Online estimation and control for autonomous electric networks

Summary

For temporally varying electric distribution networks, reliable information about the system topology and parameters may be missing or outdated. In the project we will develop online estimation algorithms for the network reconstruction, so as to automatically track changes, even in the presence of noisy and incomplete data. Furthermore, we will blend online estimation algorithms with higher-level controllers commonly used in distribution networks  and apply our methods to the moon-shot testbed developed in the NCCR. Our ultimate goal is to contribute to the development of autonomous electric networks that self-adapt in real-time to changing ambient conditions.

Online estimation and control for autonomous electric networks

For temporally varying electric distribution networks, reliable information about the system topology and parameters may be missing or outdated. In the project we will develop online estimation algorithms for the network reconstruction, so as to automatically track changes, even in the presence of noisy and incomplete data. Furthermore, we will blend online estimation algorithms with higher-level controllers commonly used in distribution networks  and apply our methods to the moon-shot testbed developed in the NCCR. Our ultimate goal is to contribute to the development of autonomous electric networks that self-adapt in real-time to changing ambient conditions.

112
cf878966-5241-4f6e-bb36-4f46dbfe98d6

Stochastic behavioural models for data-driven control

Summary

Recent results on data-based control exploit alternative parametrizations of linear and nonlinear systems, such as those provided by behavioral theory and deep neural networks. In this project we focus on Input/Output Behavioral (IOB) models that have been used for control by assuming noiseless measurement or deterministic bounded disturbances. The goal of the project is to study the relation between IOB-based control methods and the existing approaches for stochastic systems, including system identification and stochastic model predictive control. New controllers will be tested on the energy systems relevant for the NCCR.

Stochastic behavioural models for data-driven control

Recent results on data-based control exploit alternative parametrizations of linear and nonlinear systems, such as those provided by behavioral theory and deep neural networks. In this project we focus on Input/Output Behavioral (IOB) models that have been used for control by assuming noiseless measurement or deterministic bounded disturbances. The goal of the project is to study the relation between IOB-based control methods and the existing approaches for stochastic systems, including system identification and stochastic model predictive control. New controllers will be tested on the energy systems relevant for the NCCR.

111
76577dd2-3aa6-4cf9-a61a-0755301b30cb

Distributed GP learning and model-based RL

Summary

The goal of the project is to study data-driven modelling and control frameworks for large-scale systems stemming from the interconnection of several subsystems, which are of primary interest to the NCCR. For this purpose, we will develop (i) innovative probabilistic machine learning algorithms, based on Gaussian Processes (GPs), for batch and online estimation of models complying with the coupling structure of the system and (ii) model-based Distributed Reinforcement Learning (DRL) algorithms for the generation of networked control policies. Our DRL schemes will exploit the epistemic model uncertainty to drive exploration while ensuring stability of the distributed controllers.

Distributed GP learning and model-based RL

The goal of the project is to study data-driven modelling and control frameworks for large-scale systems stemming from the interconnection of several subsystems, which are of primary interest to the NCCR. For this purpose, we will develop (i) innovative probabilistic machine learning algorithms, based on Gaussian Processes (GPs), for batch and online estimation of models complying with the coupling structure of the system and (ii) model-based Distributed Reinforcement Learning (DRL) algorithms for the generation of networked control policies. Our DRL schemes will exploit the epistemic model uncertainty to drive exploration while ensuring stability of the distributed controllers.

110
63103095-1b2c-42bc-96bf-4e514b510220

Distributed Dynamic Coverage Control for On-Demand Transportation Operations

Summary

Emerging shared-mobility systems create additional opportunities to decrease car ownership and congestion. Asymmetric demand creates imbalances in the distribution of vehicles for these systems. To maximize the covered demand and decrease the waiting time of passengers, vehicle distribution has to be rebalanced with relocations. The potentially very high number of vehicles necessitates using distributed control algorithms to efficiently solve this problem. Presence of multiple companies competing to serve the same demand can be addressed via game theoretic approaches. Μatching algorithms to create shared rides can be combined with the coverage problem to increase coverage in areas with high demand. Fairness in covering low demand areas will also be investigated.

Distributed Dynamic Coverage Control for On-Demand Transportation Operations

Emerging shared-mobility systems create additional opportunities to decrease car ownership and congestion. Asymmetric demand creates imbalances in the distribution of vehicles for these systems. To maximize the covered demand and decrease the waiting time of passengers, vehicle distribution has to be rebalanced with relocations. The potentially very high number of vehicles necessitates using distributed control algorithms to efficiently solve this problem. Presence of multiple companies competing to serve the same demand can be addressed via game theoretic approaches. Μatching algorithms to create shared rides can be combined with the coverage problem to increase coverage in areas with high demand. Fairness in covering low demand areas will also be investigated.

105
91547dbb-fb8b-4801-9b24-09f41492e308