1) IDENTIFICATION AND CONTROL OF ELECTRIC DRIVES
The main subject is sensorless control of synchronous and
asynchronous electric machines. "Sensorless" means "without
position/speed sensor on the rotor shaft". Nonlinear Adaptive
Control and Internal Model Principle techniques are adopted and
combined with standard field-oriented solutions. The main
objectives are:
a) to improve performance w.r.t. existing solutions;
b) to provide formal proof of convergence and stability for the new
solutions by means of Lyapunov methods.
2) MODELLING AND CONTROL OF POWER ELECTRONICS EQUIPMENTS
Power electronics is crucial in energy generation and management
(e.g. wind turbines, smart grids). Recent developments in this
field require an integrated approach to electronic topology
definition, component dimensioning and control design. Energy-based
modelling and internal model control are exploited to tackle such
issues. Another important point is the robustness with respect to
off-nominal working conditions. In this respect, advanced
nonlinear anti-windup techniques are proposed.
3) THERMAL MODELLING AND CONTROL OF COMPUTING PLATFORMS
Upcoming multi-core and many-core systems-on-chip need a dynamical
management of running frequencies and computing task to guarantee a
safe thermal behavior. In this respect, advanced identification and
model predictive control techniques are propose to achieve compact
and scalable solutions giving reliable temperature capping, while
maximizing computing performance.
4) MODELLING AND CONTROL OF AUTOMOTIVE SYSTEMS
Past activities:
a) modelling and control of drive-line behaviour and traction by
means of servo-actuated clutch and gear-box;
b) modelling and control of cam-less systems;
c) modelling, architectural design, component sizing and
control of hybrid vehicles.
5) DESIGN METHODOLOGIES FOR LOGIC CONTROL IN AUTOMATION
Supervision and control of large-scale and distributed automation
systems require involved logic control functionalities. In the
past, some research activities have been carried out to define
design methodologies and tools where clear separation of
"mechanisms" and "policies" is exploited to obtain subsystems
composability.