The development of embedded systems is a lengthy and expensive process. That process is further complicated by the extra-functional properties of embedded systems, especially timing and energy requirements. Currently, these properties only manifest themselves towards the end of the development process, long after the most important design decisions, such as the selection of the processors, have been made. However, if it then turns out that the developed system does not meet the extra-functional requirements and, for example, reacts too slowly or drains the battery too quickly, corrections to the design are expensive if not impossible.

In TwinSpace we extend the development process of embedded systems to include early consideration of extra-functional properties. For this we use the concept of the digital resource twin. These twins are a representation of the extra-functional behavior of the target application or the target system and can be generated at the very beginning of the development process. In this manner, we can not only identify undesirable developments in the design process early on, but even optimize extra-functional properties. The aim of the University of Augsburg in the project is the automatic generation of resource twins at component level (i.e. on C code) and at system level (including communication, scheduling and middleware).