The research team has been working on several aspects of automatic control science and technology, in relation to both basic research and industrial applications. The activity is divided into four areas: Control design, Model identification and data analysis, Automation of vehicles and transportation systems, Automation of energy systems.
The research in the first area aims at developing advanced techniques for analysis, design and optimization of particular classes of dynamical systems, including switched systems, uncertain hybrid systems and cyber-physical systems. A relevant topic is the automation of large-scale systems, constituted by networks of heterogeneous and (semi-)autonomous agents, for which hierarchical and distributed predictive control methods will be developed.
The second area deals with identification and data analysis methods, based on different approaches, suitable to efficiently  solve problems of black-box or grey-box estimation, learning, filtering and prediction.
The third area covers applications in automotive, railway and  aerospace control and ranges from single vehicle control to multi-vehicle coordination and air traffic management.
The fourth area refers to modeling and control of energy systems, focusing on various topics in energy generation, distribution and management, for both conventional and innovative plants and networks.

• Distributed control and estimation. Decentralized algorithms based on the Model Predictive Control (MPC) approach have been studied, hierarchical MPC controllers for specific classes of systems have been proposed, while an innovative distributed MPC algorithm has been developed. Distributed Moving Horizon Estimators (MHE) have been designed for sensor networks, and for partitioned systems.
• Periodic control. This is a traditional area of research in our institution, in addition, the recent research activity has focused on numerical issues in the design of both time-periodic and time-invariant controllers for linear time-periodic systems.
• Control of switched systems. The main research topics within this area are: (1) design of limit cycles in oscillating systems, (2) analysis of stability, passivity and both mean-square and worst case performance under dwell-time constraints, (3) design of state-feedback and output-feedback switching laws improving H2, Hinf and passivity performance, (4) control of positive switched systems with application to mitigating HIV infection.
• Robust control. In this area the research has been concentrated on (1) blending control for simultaneous performance in input-output channels, (2) guaranteeing robustness in nonlinear Hammerstein systems with application to a dynamic engine test-bench, and (3) control of quantized systems and parameter adaptation for improving L1 robustness in nonlinear systems with application to a continuous casting problem.
• Fault-tolerant control. Design of low-order regulators which guarantee tracking and disturbance rejection in the presence of actuators and sensors faults. The considered exogenous signals have various forms and the regulators are centralized, or decentralized.
• Iterative Control. The research in this field concerned the development of a new iterative control scheme which explicitly accounts for the presence of uncertainty in the plant description (iterative robust control).
• Randomized methods. A new technique (the Scenario Approach) to efficiently solve semi-infinite optimization problem arising when dealing with uncertain systems was developed. The major achievement is a new technique which permits the modulation of robustness in robust control problems so as to avoid conservatism.
• Stochastic hybrid systems. The problem of designing a state feedback controller that maximizes the probability that a discrete time Stochastic Hybrid System (SHS) remains within a certain safety region has been addressed via dynamic programming. Computational issues involved in the numerical solution to the resulting dynamic programming equations have been studied. A two-step scheme for the approximate model checking of SHSs has been developed. A method for abstracting a SHS to simplify reachability computations was studied.
• Active control of noise and vibrations. Two methodological research topics have been addressed in this area, namely nonlinear active noise control using NARX models, to deal with microphone distortion and saturation effects and the development of efficient adaptive algorithms for the attenuation of non-Gaussian noise signals with impulsive characteristics. In addition, the implementation of a general purpose multi-tonal disturbance compensator on a high-speed computational architecture, based on a field programmable gate array (FPGA) has been studied.
• Modeling and control of discrete event systems with Petri nets. In this area, formal approaches for the modeling of flexible manufacturing systems and batch systems have been studied. In addition, several innovative deadlock prevention algorithms have been developed based on siphon control. A resource decoupling scheme is studied, whereas a set covering-based approach that optimally matches emptiable siphons to critical markings is investigated.

• Subspace Model Identification. SMI methods are now a very well established approach to deal with multivariable model identification problems. The recent research in this area has focused on the development of continuous-time subspace identification algorithms. Methods have been developed for LTI as well as for LPV models and issues related to recursive implementations have been studied.
• Interval predictor models (IPMs) identification. An IPM is a model returning confidence intervals for the prediction of future output values. A new method for the identification from experimental data of an optimal (i.e. with smallest width) IPMs with certified reliability was proposed. The main feature is that the reliability evaluation, i.e. the probability that a future output belongs to the returned interval, is performed without any assumption on the data generating system.
• Prediction error methods. In view of results showing the unreliability of the classical asymptotic theory in presence of poor information, new approaches based on re-sampling techniques (including sub-sampling, jackknife and bootstrap) was investigated.
• Nonlinear identification. The development of innovative algorithms for the identification of polynomial nonlinear ARX models has been addressed. In particular, the parameter estimation problem has been tackled by minimization of the simulation error, using prediction models with increasing horizon.
• Data-based controller design. The VRFT ("Virtual Reference Feedback Tuning") paradigm for control design is based on the reformulation of control design problems into system identification ones. VRFT allows one to directly tune a controller in a given class without performing any plant estimation procedure. The method is now ready-to-use for practical applications, since a Matlab-toolbox has also been developed.

• Road-vehicles dynamics and control. A comprehensive understanding of semi-active damping has resulted in new methods and algorithms for suspension systems.
• Single-track and narrow-track vehicles. The control of the dynamics of motorcycles and tilting 3 and 4 wheelers have been studied
• Electric vehicles. The focus of the research has been on the co-design of vehicle-dynamics and energy-management control strategies. Two new vehicles have been conceived, patented and prototyped: the "Bike+" (a new concept of non-plug-in electric bike) and the "Flyboard" (a vehicle for indoor mobility with smartphone-based HMI).
• Lithium-batteries modelling and control. The research activity has been focused on the development of new models for various types of Li-ion and Li-po cells. The models has been used for the development of State-of-Charge estimation and State-of-Health estimation algorithms. The research has been conducted both using test-bench and on-vehicle experiments.
• Intelligent transport systems. In this field the main results have been obtained in two directions: the development of real-time algorithms for the estimation of the drive-style, and the development of an innovative intersection-support architecture, based on smartphones and on a centralized algorithm.
• Active control of helicopter dynamics. The research has focused on the analysis of coupled rotor/fuselage dynamics under the effect of active control schemes such as Higher Harmonic Control for vibration attenuation. A discrete-time approach has been proposed, taking into account periodicity (both of the rotor and of the controller) in analysing the stability of the closed-loop system. The activity has been developed in cooperation with the University of Maryland Rotorcraft Center.
• Spacecraft dynamics and control. Various problems associated with estimation and control of spacecraft dynamics have been considered. Among others: modeling and simulation tools for spacecraft dynamics have been developed, in an object-oriented framework, in cooperation with DLR (German Aerospace Center); a novel algorithm for attitude estimation on the basis vector measurements has been proposed.
• Air traffic management systems. A methodology was developed for the reachability analysis of stochastic hybrid systems with application to aircraft conflict prediction. In addition, the characterization of new metrics for the assessment of complexity in air traffic, suitable for new generation air traffic management systems with autonomous aircraft, has been addressed. The research team obtained the leadership of the work-package on “Prediction of complex traffic conditions” under the FP6 EC funded STREP project IFLY.

• Modeling and control of floating wind turbines. The research has focused on the development of low order physical models of floating wind turbines with Tension Leg Platforms and Spark Buoy platforms. These models have been used for the design and testing of multivariable regulators with the Hinf approach. The control systems have been validated in simulation on reference simulators widely used in the international research community.
• Voltage control of smart grids. An accurate dynamic simulation environment has been developed according to an object-oriented paradigm. This tool has been used to design distributed control systems for voltage control in medium voltage radial feeders. The control scheme is made by local PI regulators for control of the distributed generators and by one regulator for each feeder, either PI or MPC.
• Energy systems modeling and control. In this field the research focused on innovative techniques for the attenuation of carbon dioxide in flues gases, fuel cells control, smart-grid management.
• Transport-related energy-management. In this field the research activity has been focused on the development of the estimation of the impact on the grid of large fleets of electric vehicles, and on the joint use of the vehicle battery-packs for mobility and for grid energy-storage.