The project is aimed at contributing new methods and techniques for developing cyber- physical systems on multicore platforms with support for running applications with temporal restrictions, such as control systems, and multiple criticality levels. The project addresses both the definition of the layers of software that make up the architecture of the system (including the execution platform), and the methods and tools for designing, analysing, and deploying the system. It is based on a model-driven engineering approach, and aims to develop systems with subsystems or applications of different levels of criticality. Ways of facilitating the eventual individual certification of the subsystems that may require it, will be explored.

The proposed approach is based on the concept of using virtual machines to build a partitioned architecture, so that the subsystems of different levels of criticality can run on different virtual machines or partitions. The hypervisor that implements this concept must ensure isolation between partitions in both the time and storage space domains, in order to ensure that any errors in the less critical components are not propagated to the most critical ones nor disrupt their proper execution.

The extension of this concept to multicore platforms is a fundamental contribution of this project. The interferences that arise in this kind of systems when accessing shared hardware resources hinder isolation between processes running on different cores. The different types of interference will be analysed in the project, and the most appropriate techniques for compensating for it, as well as the most suitable scheduling methods for such systems will be investigated.

The project will also address the methodological issues associated with modelling, design and implementation of cyber-physical systems. A prototype of an integrated framework with tools for analysis and design of systems will be developed. The control aspects are very important in cyber-physical system. Consequently, a special effort will be devoted to analysing their impact on the design, and specific tools will be developed to this purpose.

Project results will be applied to two case studies in orde r to verify their applicability in industry. A remotely piloted aircraft system (RPAS), and the embedded on-board software for the university satellite UPMSat2 have been selected.