Cyber-physical Systems of Systems (CPSoS) are large complex systems where physical elements interact with and are controlled by a large number of distributed and networked computing elements as well as human users. Their increasingly stringent demands on efficient use of resources, high service and product quality levels and, of course low cost and competitiveness on the world market introduce big challenges related to the design operation continuum of dependable connected CPSs. The CPSoSaware project aims at developing the models and software tools to allocate computational power/resources to the CPS end devices and autonomously determining what cyber-physical processes will be handled by the devices' heterogeneous components (CPUs, GPUs, FPGA fabric, software stacks). The project relies on Artificial Intelligence (AI) support to strengthen reliability, fault tolerance and security at system level and also to lead to CPS designs that work in a decentralized way, collaboratively, in an equilibrium, by sharing tasks and data with minimal central intervention. The CPSoSaware system will interact with the human users/operators through extended reality visual and touchable interfaces increasing situational awareness. The CPSoSaware system will be evaluated: i) in the automotive sector, in mixed traffic environments with semi-autonomous connected vehicles and ii) in the manufacturing industry where inspection and repair scenarios are employed using collaborative robots.