This line of activity aims at offering comprehensive expertise about most aspects of embedded system design and development for academic and industrial applied research alike. It covers both lower-level architectural topics—mainly related to operating system design, implementation, and performance assessment—and higher-level design and verification tools and techniques, since both are needed to support embedded systems of today and tomorrow.
Multicore Architecture for Real Time Processing
Multicore processors used to be an answer to the data processing demands of specialized devices, like packet-level network processors, but are nowadays ubiquitous in desktop computing and are becoming extremely popular in many other application domains, ranging from mobile phones to hard real-time systems. Along with obvious benefits, mainly related to reduced power consumption and increased throughput, multicore architectures are bringing epocal changes to multiple areas of computer science and engineering, encompassing operating systems, task scheduling, inter-process communication, and application design. In this domain, our main goal is to bring the multicore revolution not only to mainstream, mass-produced appliances, but also to custom industrial embedded systems. Although those systems could draw significant benefits from multicores, the lack of concrete design information and use cases, along with stringent dependability requirements, have so far slowed down innovation.
Model-based Software Engineering
Software systems are nowadays becoming more and more complex and are subject to an ever-increasing number of requirements in diverse areas. This trend puts a strain on traditional design processes, which could scarcely cope with the added complexity, and has recently led to the introduction of more advanced methods rooted in formal analysis and verification. Model-Based Software Engineering (MBSE) is an enabling technique towards this direction and is steadily gaining acceptance in Cyber-Physical Systems (CPS) design. Using models as main artifacts brings key advantages like offering an always up-to-date formal system specification, as well as the ability to perform early-stage validation and to generate code from the models. Research activities are targeted at bridging the gap between MBSE and non-functional constraints related to real-time execution, energy consumption, security, and dependability, which have become essential in CPS design but were traditionally overlooked by MBSE.