scope:

The book contains 11 chapters.Chapter 1 provides a general introduction to the addressed matter and motivates the adopted approach, also by means of a very synthetic historical analysis. Chapter 2 introduces the few required basic elements of systems theory, as anticipated. This chapter and the following Chapter 4 end with a `problems' section, to allow the interested reader to verify their acquisition of the illustrated concepts. Chapter 3 applies the modelling-related ideas of Chapter 2 to the particular domain addressed herein, evidencing its peculiarities. Some introductory examples are reported and commented on. Chapter 4 deals with the required basic elements of control theory, adopting an attitude, and consequently an organisation, analogous to those of Chapter 2. Chapter 5 deals with task scheduling. A general dynamic model is proposed, and based on that, a methodology is presented to design a scheduler along the control based paradigm. Said methodology is then applied, leading to two scheduling algorithms, that are comparatively tested with respect to classical (non-controlbased) ones. Chapter 6 addresses the problem of memory management, and proposes for it too a control-based solution. Simulation results are presented and commented on, to evidence the obtained advantages. Chapter 7 presents, limiting again the scope to the bare essential, some control techniques - more advanced than those of Chapter 4 - that are used in the following one. In particular, the chapter deals with Model Predictive Control and model identification, and ends with a `problems' section for the same reason as Chapters 2 and 4. Chapter 8 treats more in general the resource allocation problem. Thanks to the adoption of more a high-level viewpoint with respect to the previous chapters, here a design methodology is sketched out and supported, that can be used as a modus operandi when addressing numerous problems that appear quite different from the application point of view, but in the context of the systems theory have a definitely uniform mathematical structure. Chapter 9 applies the ideas presented so far to power-aware resource management. Apart from discussing another interesting application case, the main point here is to evidence how a control-based design attitude is naturally keen to host problems with very different requirements. Chapter 10 presents the Miosix kernel, i.e., the nucleus of an operating system that is being developed along the proposed approach. Motivations for developing an ad hoc kernel are provided, the realised functionalities are described, and future developments are outlined. Chapter 11, after all the different issues touched in the previous ones, suggests a way of casting them into a unitary view conceptually connected to that of `cyberphysical'systems, thereby discussing - as a work in progress - some further ways in which a control-theoretical point of view can be helpful for design, analysis and assessment.

Authors

Alberto Leva, Martina Maggio, Alessandro Vittorio Papadopoulos and Federico Terraneo