Design techniques for high-reliability complex electronic systems

As the market continuously requires ever more powerful systems, there is a huge economic incentive in technological advances that can make that required power available to the end user of the system. Conversely, more powerful devices make it possible to integrate a large number of functions inside a single system. Eventually, users want to assign mission- critical or safety-critical tasks to digital systems. This kind of delegation makes reliability a major requirement for digital systems. The activities to achieve system reliability include validation, verification, testing of the single components and of the entire system, diagnosis of the components, system hardening. These activities in turn benefit from the development of methodologies and tools for their effective application. The performed activity concentrates on test and diagnosis of digital programmable circuits, with particular emphasis on microprocessor cores and micrcontrollers. In par- ticular, the focus of activity has been the development of approximate methodologies for circuit test and diagnosis, and the assessment of those methodologies on realistic case studies. The activity also touched the field of system hardening, with the development of a methodology having the property of being transparent. As a support activity for the test and diagnosis field, work has been performed in the field of evolutionary computation. The work includes the development of evolutionary methodologies, the definition of tool architecture, the coding, total or partial, of such tools, and the application of evolutionary approaches to problems seemingly far from the CAD field, with the purpose of enhancing the tools performance. The thesis is structured as follows. Chapter 1 provides a very brief introduction to the fields of test, diagnosis, system hardening and evolutionary computation, also defining the terminology for the following. In Chapter 2 the activities performed in the field of test are presented. In chapter 3 the activities pertaining to diagnosis are described. Chapter 4 details the activity performed for system hardening. Finally, chapter 5 describes the diverse activities performed in the field of evolutionary computation.