Model-driven quality and resource management for CPSs

A Cyber-Physical System (CPS) integrates cyber systems, human users, networks and physical systems. Thus, a CPS needs visual context and awareness to make autonomous and correct decisions. Advanced image and video processing is computationally intensive and challenging. Moreover, a CPS comprises increasingly complex and distributed configurations, which is reflected
in the growing number of sensors, actuators and other smart devices. This leads
to an exponential number of dynamic system configurations. To make matters
worse, a CPS needs to simultaneously satisfy many rigorous constraints, e.g.,
hard deadlines, safety, quality, and performance. Hence, the system designer
is confronted with an immense number of potential configurations of which a
number meet the constraints and only a fraction are optimal regarding certain
qualities. This makes finding the optimal configurations hard, especially during
run-time. A domain-specific language (DSL) for quality and resource managment (QRM) is presented to specify these configurations conveniently and reason
about them in an automated manner.