RICERCA ITALIANA     

Advanced control methodologies for hybrid dynamical systems

Università degli Studi di Cagliari

Abstract

The main objective of this project is to bring together five Italian research units that are actively working in the area of Hybrid Systems (HS). HS are now a very hot domain in control theory, as they are becoming the scientific paradigm to systematically address the problem of analyzing, modelling, simulating, synthesizing, and optimizing embedded systems, where digital controllers interact with physical plants. The project will focus the common research activity towards a single goal: derive new advanced methodologies for the control of HS, with a particular emphasis on the broad family of switched systems models.

The project is structured in five technical workpackages (WP) that cover the main themes of interest in the area of control of HS. Each WP is further structured into activities devoted to special classes of models or to particular aspects of the problem tackled in the corresponding workpackage.

WP1 is devoted to the use of optimization techniques in control. This approach has proved to be extremely fruitful for the control of a wide class of problems that can be modeled by HS, but several problems are still open and need to be dealt with. The WP contains three activities dealing with different hybrid models that the optimization is based on: time-driven model, event-driven models and stochastic models.

WP2 deals with optimal control of HS. Optimal control is a classical problem, yet the research community is still >>>

Principal Investigator

Alessandro GIUA Università degli Studi di CAGLIARI

Research Objectives

Recent technological innovations have caused an ever increasing interest in the study of hybrid systems (HS). The peculiarity of hybrid systems is the interaction between continuous-time dynamics (governed by differential or difference equations), and discrete dynamics and logic rules (described by temporal logic, finite state machines, if-then-else conditions, discrete events, etc.).

The research domain of HS is attracting an ever increasing number of researches, as shown in the last few years by the number of sessions dedicated to this topic in the most important automatic control conferences (such as the IEEE Conference on Decision and Control, or the IFAC World Congress), and by the number of relevant papers that have been published on leading journals (such as the IEEE Trans. on Automatic Control, or Automatica). The growing interest by both academia and industry in hybrid systems is motivated not only by theoretical challenges, but mainly by their ability to model, analyse and synthesize controllers in a large variety of application areas.

However, hybrid dynamics are often so complex that a satisfactory feedback controller cannot be synthesized by using analytical tools similar to those that pertain to linear systems or to certain classes of nonlinear smooth systems. Heuristic design procedures usually require trial and error sessions, extensive testing, are time consuming, costly and often inadequate to deal with the complexity of the >>>

Timescale

24 months

National and international background

The peculiarity of hybrid systems is the interaction between continuous-time dynamics (governed by differential or difference equations), and discrete dynamics and logic rules (described by temporal logic, finite state machines, if-then-else conditions, discrete events, etc.). Several modelling frameworks for hybrid systems have appeared in the literature [1,2,3]: each class of hybrid systems is usually tailored to solve a particular problem, and many models look largely dissimilar. At the present stage of development of the field, this plethora of models is unavoidable. There is, however, a family of models that has received a lot of attention as witnessed by the hundreds of scientific papers and by the recent research books [4,5] devoted to this topic: this broad class is called "switched systems" and can be considered in a way as a basic model. This justifies the particular attention devoted in this project to this family.

The growing interest by both academia and industry in hybrid systems is motivated not only by theoretical challenges, but mainly by their ability to model, analyse and synthesize controllers in a large variety of application areas, including automotive systems [7] and TCP/IP network problems [8]. A feedback controller for a hybrid dynamical system is an on-line automatic decision-making algorithm that repeatedly observes the environment and reacts to it with a proper decision, in order to meet the specifications (such as quality >>>